Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Methodology for analysis and reporting patterns of failure in the Era of IMRT: head and neck cancer applications

Methodology for analysis and reporting patterns of failure in the Era of IMRT: head and neck... Background: The aim of this study is to develop a methodology to standardize the analysis and reporting of the patterns of loco-regional failure after IMRT of head and neck cancer. Material and Methods: Twenty-one patients with evidence of local and/or regional failure following IMRT for head-and-neck cancer were retrospectively reviewed under approved IRB protocol. Manually delineated recurrent gross disease (rGTV) on the diagnostic CT documenting recurrence (rCT) was co-registered with the original planning CT (pCT) using both deformable (DIR) and rigid (RIR) image registration software. Subsequently, mapped rGTVs were compared relative to original planning target volumes (TVs) and dose using a centroid-based approaches. Failures were then classified into five types based on combined spatial and dosimetric criteria; A (central high dose), B (peripheral high dose), C (central elective dose), D (peripheral elective dose), and E (extraneous dose). Results: A total of 26 recurrences were identified. Using DIR, recurrences were assigned to more central TVs compared to RIR as detected using the spatial centroid-based method (p = 0.0002). rGTVs mapped using DIR had statistically significant higher mean doses when compared to rGTVs mapped rigidly (mean dose 70 vs. 69 Gy, p =0. 03). According to the proposed classification 22 out of 26 failures were of type A (central high dose) as assessed by DIR method compared to 18 out of 26 for the RIR because of the tendencey of RIR to assign failures more peripherally. Conclusions: RIR tends to assigns failures more peripherally. DIR-based methods showed that the vast majority of failures originated in the high dose target volumes and received full prescribed doses suggesting biological rather than technology-related causes of failure. Validated DIR-based registration is recommended for accurate failure characterization and a novel typology-indicative taxonomy is recommended for failure reporting in the IMRT era. Keywords: Patterns of failure, IMRT, Head-and-neck cancer, Deformable image registration Introduction (defined as tumor persistence or recurrence) including in- Intensity-modulated radiation therapy (IMRT) is one of adequate definition of the tumor extension and clinically the most important innovations in modern radiation ther- important target volumes (TVs), uncertainties related to apy and represents a paradigm shift in the treatment of daily positioning, weight loss or deformation of tumor and head and neck cancers (HNCs). However, there are certain normal tissues during the course of treatment, and uncer- hazards that may increase the risk of loco-regional failure tainties in plan optimization, dose calculation and treat- ment delivery [1–5]. The accurate and specific definition of the exact site of * Correspondence: asmohamed@mdanderson.org; cdfuller@mdanderson.org failure, in addition to the radiation dose given to this site Head and Neck Section, Division of Radiation Oncology, Department of is, therefore, mandatory to identify the possible cause(s) Radiation Oncology, The University of Texas M.D. Anderson Cancer Center, Box 00971515 Holcombe Blvd, Houston, TX 77030, USA of failure. The classic definition of failures as “local”,or Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Mohamed et al. Radiation Oncology (2016) 11:95 Page 2 of 10 “regional”, was appropriate in the setting of conventional documented recurrence. A total of 21 cases were ran- radiotherapy using large homogeneous dose-volumes, domly selected from the recurrence dataset based on the but is no longer helpful nor descriptive of distinct types following eligibility criteria: IMRT given for curative in- of failure in patients treated with IMRT [6–8]. tent; treatment of intact tumor (i.e. post-operative cases Several previous efforts have addressed the importance were excluded); equal distribution of various head and of studying the patterns of failure after IMRT treatment of neck subsites (i.e. nasopharynx, oropharynx, hypopharynx, HNCs, [2, 4, 9–13] with most reporting failures as and lateral neck “i.e. neck nodes of unknown primary “infield”, “marginal” or “outfield” based on the percentage site”); radiological evidence of local and/or regional failure; of overlap between the failure volume and the respective available CT scan of failure site prior to any salvage ther- TV on the treatment planning CT (pCT) [4, 9, 10, 12, 13]. apy; and pathologic and/or radiologic evidence of recur- The ability to accurately describe the relation of failure rence (i.e. biopsy, or high SUV on PET). to original TVs and dose mandates a fairly precise method to co-register the diagnostic CT documenting IMRT treatment planning and delivery recurrence (rCT) to the original pCT. However, the ma- All patients had been positioned supine in an individualized jority of the previous studies implemented mainly rigid thermoplastic head and shoulder mask for CT simulation image registration techniques (RIR) [2, 4, 10, 12, 13]. and treatment and a custom dental stent used as an RIR is simple, quick and widely used but it allows only intraoral immobilization device [19, 20]. A treatment for 6° of freedom and doesn’t account for changes in the pCT scan was used for defining TVs. Target volume shapes or relative positions of different regions-of- definition was done in Pinnacle treatment planning interests (ROIs) [14]. Emerging data demonstrate the system (Pinnacle, Phillips Medical Systems, Andover, superiority of deformable image registration (DIR) com- MA), with rigorous multi-physician target delineation pared to RIR in registering pCT to on-treatment CT or and quality assurance [21, 22]. conebeam CT in the setting of image guided radiother- Treatment was uniformly delivered using Varian (Varian apy (IGRT) for HNCs [15–17]. However, very few stud- Medical Systems, Palo Alto, CA) linear accelerators ies addressed DIR software implementation for the delivering 6-MV photons. Three clinical target volumes purpose of registering the diagnostic rCT to the original (CTVs) were typically defined. CTV definitions and dose pCT [6, 7]. prescriptions are summarized in Additional file 1: Our group has recently validated different registration Table S1. Treatment was delivered in a conventional frac- techniques used for co-registering diagnostic contrast en- tionation scheme (average 33 fractions). Patients were hanced head and neck CT to non-contrast planning CT treated using a monoisocentric technique with an antero- and showed DIR was superior for this application [18]. As posterior low-neck supraclavicular field matched to the a continuation of these efforts and to validate DIR as a IMRT fields or using whole neck IMRT for cases where tool to improve accurate definition of the patterns of loco- gross nodes are located at the match line. regional failure in the era of IMRT for HNCs, we sought to undergo the following specific aims: Post-treatment follow up Initial post-treatment evaluations were made at 8–12 1) Develop a workflow methodology to standardize the weeks after therapy completion and subsequently every analysis of HNCs patterns of failure using both 2–3 months for the first year, every 3–4 months for the geometric and dosimetric parameters. second year, and at least twice a year up to 5 years. 2) Assess the impact of registration (rigid vs. deformable) techniques on patterns of failure Loco-regional failure quantitative analytic parameters. Cases where local and/or regional recurrence was re- 3) Develop a granular classification and nomenclature corded had their immediate post-failure diagnostic im- to optimize the accurate reporting of distinct failure ages exported as DICOM files from the clinical PACS typology. system to the treatment planning system, where radio- logical evident recurrent gross disease (rGTV) was Material and methods manually contoured by a radiation oncologist (ASRM) Patients and reviewed by a head and neck service-specific attend- Tumor registry data for patients diagnosed with head ing radiation oncologist (CDF). and neck squamous cell carcinoma, whom were treated by IMRT at The University of Texas, MD Anderson Image registration Cancer Center between 2006 and 2009, were retrospect- For each patient, the rCT or rPET-CT was co-registered ively reviewed under an institutional review board ap- with pCT using both rigid and deformable image registra- proval. 600 patients were identified, of those 103 had a tion techniques. DIR was performed using a commercial Mohamed et al. Radiation Oncology (2016) 11:95 Page 3 of 10 software (ADMIRE, Elekta AB, 2015) validated previously the Pinnacle planning system to rigidly align rCT to by our group for the registration of contrast-enhanced pCT, following that rGTVs where exported to patient’s diagnostic CT to non-contrast enhanced planning CT plan where dose volume histograms (DVHs) and rGTV [18]. Deformation vector fields were obtained from DIR centroids were generated and analysis metrics were cal- algorithm, mapping the deformation of the rCT onto the culated. Figure 1 illustrates the workflow process of our pCT. Subsequently, in a custom written Matlab routine registration methodology. (MATLAB R2013a, The MathWorks Inc., Natick, MA, 2013), pCT; dose grid; original plan TVs; rCT; and rGTVs Analysis of failure metrics were imported. The deformation fields were then applied For both RIR and DIR mapped rGTVs the following to rGTVs segmented on the rCT to convert them into ‘de- metrics were evaluated: formed rGTVs’ on the pCT. 1) Recurrence volume, 2) Location of the centroid Evaluation of deformed rGTVs relative to original relative to planning TV: Centroid is the central voxel of planning TVs was done using both centroid-based the recurrence volume plus added 2 mm margin to ac- method that assumed the center of mass of rGTV was count for registration error, 3) Spatial relationship of the origin of the recurrence volume and its location was rGTV centroids to IMRT/supraclavicular match line and compared relative to planning TV after applying deform- ipsilateral parotid in case of peri-parotid failure, and 4) ation vector fields (DVF). Simultaneously, RIR was per- Mean and maximum dose to rGTVs, dose to 95 % fail- formed using the rigid co-registration tool available in ure volume (fD95%), and mean dose to centroid volume. Fig. 1 Workflow process of patterns of loco-regional failure registration process Mohamed et al. Radiation Oncology (2016) 11:95 Page 4 of 10 Classification of failure  Type D: Peripheral elective dose failure, where the In order to refine our reporting and quality assurance failure centroid originates from elective dose TV and practices using a standard nomenclature, we developed a its fD95% receives <95 % dose prescribed to the granular typology of failure categories relative to the respective TV. planning TV and dose. As illustrated in Fig. 2, failures  Type E: Extraneous dose failure, where the failure were classified into five types based on combined spatial centroid originates outside all TVs. and dosimetric criteria: For patients treated with low-neck supraclavicular field Type A: Central high dose failure, where the mapped matched to the IMRT fields, two additional types were failure centroid originates in high dose TV and the added: dose to 95 % failure volume (fD95%) is ≥95 % dose prescribed to corresponding TV of origin.  Type F: Junctional failures at the site of IMRT/ Type B: Peripheral high dose failure, where the supraclavicular match line. mapped failure centroid originates from high dose  Type G: Low neck failures at the site of low-neck TV but its fD95% receives <95 % dose prescribed to supraclavicular field. this TV. Type C: Central elective dose failure, where the Statistical analysis failure centroid originates from elective dose TV Non parametric statistics were used to compare analysis and its fD95% receives ≥95 % dose prescribed to the metrics for centroid locations and dosimetric parameters respective TV. of failures mapped using RIR versus DIR registration Fig. 2 Classification scheme of IMRT patterns of failure using combined centroid based geometric method coupled with the dosimetric parameters. Panel a) shows an example of types A (central high dose) and C (central elective dose) failures, panel b) shows an example of types B (peripheral high dose) and D (peripheral elective dose) failures, and panel c) shows an example of type E (extraneous dose) failure. Mohamed et al. Radiation Oncology (2016) 11:95 Page 5 of 10 techniques. A p-value ≤ 0.05 was deemed significant. Table 1 Patient demographics, disease, and treatment characteristics Statistical assessment and data tabulation was performed using JMP v 11Pro (SAS institute, Cary, NC). Total n = 21 (%) Results Age (years) Patients Median 58 A total of 21 patients with HNSCC were included in this Range 30–75 pilot methodology/workflow development study. Median Time to Failure (months) age was 58 years (range 30–75), and 86 % were men. Median 12 Patient, disease, and treatment characteristics are pre- sented in Table 1. Recurrences were delineated using Range 5–69 diagnostic contrast-enhanced CT in 16 patients and Sex using PET-CT in 5 patients. Male 18 (86) Female 3 (14) Spatial/dosimetric failure mapping Origin Spatial mapping Nasopharynx 6 (28) A total number of 26 rGTVs were delineated. Mean rGTVs volume was 12.5 cm (range 1–105). The regis- Oropharynx 5 (24) tration method independently affected the spatial loca- Hypopharynx 5 (24) tion of mapped failures. Failures mapped using DIR Unknown primary 5 (24) were significantly assigned to more central TVs com- T-category pared to failures mapped using RIR. 38 % of centroids T0 5 (24) (n = 10) mapped using RIR were located peripheral to T1 1 (5) the same centroids mapped using DIR (p = 0.0002). Table 2 illustrates the sites and geometric details of all T2 7 (33) failures mapped to the pCT. T3 5 (24) T4 3 (14) Dosimetric mapping N-category rGTVs mapped using DIR had statistically significant N0 1 (5) higher mean doses when compared to rGTVs mapped N1 5 (24) rigidly (mean dose 70 vs. 69 Gy, p = 0.03) while compari- son of mean fD95% was not statistically significant N2 12 (57) (mean fD95% 68 vs. 66 Gy, p = 0.07), and comparison of N3 3 (14) maximum, and centroid doses showed no-significant dif- Treatment ferences between both registration methods (p = 0.7 and Radiation alone 4 (19) 0.4, respectively). Additional file 1: Table S2 shows the Concurrent ChemoRadiation 9 (43) dosimetric details of all failures. Induction Chemotherapy + Radiation 1 (5) Classification of failure Induction Chemotherapy + Concurrent ChemoRadiation 7 (33) Based on the proposed classification of failure using Radiation dose both the spatial location of the centroids of the mapped Mean (SD) 69.2 (1.7) failure volumes coupled with the dosimetric parameters Radiation fractions (as illustrated in Fig. 2), 22 (84.6 %) out of the 26 failures Mean (SD) 33 (2) mapped using DIR were of type A, one of type B, 2 of type C, and one of type G. Whereas, 18 (69 %) out of the 26 failure mapped using RIR were of type A, 5 of located peripheral to the same centroids mapped using type B, 2 of type C and one of type G. Figure 3 illustrates DIR as shown in Table 2. However, after adding the the difference in classification using both registration dosimetric component of analysis, only 4 of those 10 methods. There was no type F (junctional) failures in pa- RIR mapped rGTVs were peripheral high dose failures tient subset treated using anteroposterior low-neck (type B) and the other 6 were central high dose failure supraclavicular field matched to the IMRT fields. Addition- (type A) because despite the centroids were spatially per- ally, no peri-parotid failures were detected. ipheral in location to the respective DIR ones but dosi- This combined spatial/dosimetric analysis shows that metrically, the rGTVs 95 % volumes still had ≥95 % while 10 centroids (38 %) of RIR mapped rGTVs were dose. Figure 4 shows an example of the differences in Mohamed et al. Radiation Oncology (2016) 11:95 Page 6 of 10 Table 2 Geometric details of failed rGTVs reference. However, such a reporting language gives no information regarding radiation fields/volumes, or deliv- n. Percent ered dose. In the pre-IMRT era, when large fields of N. of recurrences 26 homogenous dose were used, the definition of “in-field” Recurrence volume failure (i.e. within the field borders) or “marginal” failure Mean (SD) 12.5 (23) (i.e. adjacent to the block edges) were intuitive descrip- Location of centroid using RIR tors relating treatment parameters to sites of failure. GTV 12 (46) However, in the current era of conformal therapy [23, 24], CTV1 7 (27) dose gradients, and multiple TVs make relating spatially accurate information about dose and recurrence far more CTV2 1 (4) complicated for IMRT plans. In the same way that a stan- CTV3 1 (4) dardized method for analysis and reporting for TVs has PTV1 4 (15) been undertaken successfully [23–25], a similar effort is Supraclavicular field 1 (4) desirable for pattern of failure reporting. In our opinion, Location of centroid using DIR reporting failure using only anatomic/field referents is in- GTV 22 (84) sufficient for complex multi-volume/multi-dose plans, and obscures clinically useful information which might CTV1 1 (4) lead to improvements in future studies. CTV2 1 (4) Likewise, the requirement of rigorous assurance for CTV3 1 (4) correctly localizing disease after-therapy is increased in Supraclavicular field 1 (4) terms of required spatial accuracy. The steep dose gradi- Abbreviations: DIR Deformable image registration, RR Rigid Registration, GTV ents of modern IMRT plans and proximate transition gross tumor volume, CTV clinical target volume, PTV planning target volume from high-risk CTVs to intermediate- or lower-risk CTVs implies inaccurate registration will erroneously as- spatial and dosimetric parameters for a DIR versus RIR sign the location of failure to incorrect dose/prescription mapped failure. Those 4 rGTVs were seen in the follow- volume. We, in fact, show RIR for IMRT plans resulted ing patients: two nasopharyngeal (one primary “Fig. 4” in incorrect localization relative to prior TVs and dose and one nodal site); one oropharyngeal (primary site); in 16 % of failures in the current pilot dataset. Conse- and one unkown primary (nodal site). The secondary quently, this study presents a methodology and workflow qualitative review by expert radiation oncologists (CDF, that involves the application of quality assured DIR soft- DIR) of those 4 patients agreed with DIR classification ware as a tool to standardize co-registration and to cor- that those recurrences are actually central rather than rectly attribute sites of loco-regional failure. peripheral in origin. Almost all previous studies have used RIR to de- scribe the patterns of loco-regional failure after IMRT Discussion [2, 10, 12, 13]. Chao et al. [10] reported 17/126 loco- Traditionally, failure reporting for HNCs has simply regional failures treated by definitive or postoperative classified disease as “local”, “regional”,or “locoregional”, IMRT; 53 % of failures were inside CTV1, 12 % mar- thus relating location of failure to a crude anatomic ginal to CTV1, 6 % marginal to CTV2, and 28 % Fig. 3 Bar chart illustrating the difference in failure classification using rigid (RIR) vs. deformable (DIR) image registration methods Mohamed et al. Radiation Oncology (2016) 11:95 Page 7 of 10 Fig. 4 A case of T2N0 Nasopharyngeal carcinoma recurred 63 months after IMRT. The upper panel shows the axial, coronal and sagittal images of a RIR mapped rGTV on the original pCT where its centroid is located at CTV1and the 95 % rGTV volume contained on more peripheral PTV2 (contour not shown). The middle panel shows DIR mapped rGTV on the original pCT where its centroid located at GTV and the 95 % rGTV volume contained on more peripheral CTV2. The lower panel shows RIR and DIR mapped rGTVs overlaid to plan isodose line. Note that RIR rGTV fD95% extends beyond the 95 % isodose line “66.5 Gy” (red arrow in sagittal image) which would erroneously characterize it as type B failure, while in fact DIR shows it as a type A failure (i.e. the fD95% of DIR mapped rGTV is completely encapsulated with 95 % isodose line, shown by white arrow in sagittal image) were out of field. Eisbruch et al. [2] reported on 21 RIR. Our previous work [18], as well as the current recurrences in 133 patients with non-nasopharyngeal study, confirm the qualitative superiority, in HNC ap- HNCs treated with parotid-sparing IMRT; 17/21 were plications, of DIR for CT-CT registration. in-field. Daly et al. [12] reported on 69 HNCs treated In our classification scheme, we designed a combined with parotid-sparing IMRT; 8 patients developed a geometric/dosimetric typology definition to avoid the loco-regional failure, 7 relapsed within the high-dose drawbacks of using each method separately. Centroid CTV, with one junctional failure observed. Sanguineti only method suppose a single point of origin and ignore et al. [13] described the patterns of failure for 50 pa- the dose given to the whole area of recurrence while the tients with IMRT for oropharyngeal SCC; 5 recur- dosimetric only analysis is agnostic to the geometric re- rences were related to high dose regions while 4 were currence origin. Due et al. [7] previously reported that at the low dose regions. All these reports relied on focal methods, such as the centroid method we used, are RIR, known to be less spatially accurate than DIR more accurate to localize the origins of loco-regional re- [18]; it is conceivable these results might be altered if currences than volume overlap methods, which may in- DIR methods were used. Due et al. [7] reported that correctly assign recurrences to more peripheral TVs. DIR showed slightly better reproducibility in identifi- Raktoe et al. [8] further confirmed the superiority of cation of the site of recurrence origin compared to focal methods like centroid expansion to the volumetric Mohamed et al. Radiation Oncology (2016) 11:95 Page 8 of 10 method in identifying the origin of loco-regional recur- those recurrences are actually central rather than per- rences. The combined method we used identify the esti- ipheral in origin. mated site of recurrence origin relative to the respective In this limited pilot dataset, our results showed the TV in the planning CT and then compare the dose to majority (84.6 %) of DIR mapped failures were of type A the mapped recurrence volume with the dose prescribed indicating, that biological, non-technically/non-operator to the TV of origin. Using this method, our results dependant explanations for failure predominated. How- showed that DIR significantly assigned failures to more ever, using RIR type A failures would have been errone- central TVs and doses compared to RIR concordant with ously reported as comprising only 69 %. These results the results of Due et al. [7]. assert the need for a robust, quality assured image regis- Our proposed nomenclature allows granular reporting tration technique, as error in the registration process of different types of failure. In our classification, type A would invalidate subsequent results and thus might “central high dose” failures, are considered to be bio- deceptively indicate a greater rate of technical/operator- logical failures, as they likely represent resistance to attributable therapy failures than DIR demonstrates. The maximal therapy, and thus could not conceivably be pre- current study, while underpowered to make clinical ex- vented by technical/operator dependant processes in- trapolations due to limited number of patients, nonethe- cluding IMRT QA or delineation alteration. Type A less serves as a benchmark to describe our standardized failures motivate future investigation of alternative treat- analytic and reporting method. Already, RTOG 1216, for ment stratgies (e.g. integration of novel targeted drug example, contains provisions regarding collection of im- therapies or dose escalation to identifiable biologically aging data post-failure [26], which will allow careful ana- aggressive subvolumes). Likewise, type E “extraneous lysis, and process quality improvement for future trials dose” failures cannot be modified by IMRT QA pro- and large scale datasets. cesses. They represent a possible diagnostic or decision error rather than a target delineation error (i.e. “One will Conclusions never hit what one does not aim at.”). However, type B, Rigid image registration method tends to assigns failures C, and D failures are of a special concern since they more peripherally compared with deformable method. entail potential technical or radiotherapy process fail- Using DIR, the vast majority of failures in the presented ures. Type C “central elective (intermediate or low) pilot study originated in the high dose target volumes dose” failures may be prevented by prescribing higher and received full prescribed doses suggesting biological doses (i.e. shifting to higher CTV levels). Importantly, rather than technology-related causes of failure. We type B “peripheral high dose” or D “peripheral elective heavily recommend a validated DIR-based registration dose” failures necessitate a rigorous QA process includ- technique in addition to granular combined geometric- ing triple DIR registration of pre-therapy diagnostic im- and dosimetric-based failure characterization using aging (diagnostic CT, MRI, and/or PET-CT) to pCT and novel typology-indicative taxonomy as a standard part of the earliest rCT, to assess the potential causes: potential large-scale patterns of failure reporting in the IMRT era. target delineation or dose delivery error (modifiable) ver- sus overgrown recurrence that represents actual type A Additional file or C failure which is converted to type B or D, respect- ively, due to rapidly progressive disease or neglected late Additional file 1: Table S1. IMRT target volume definitions and dose prescription. Table S2. Dosimetric patterns of failure. (DOCX 16 kb) diagnosed recurrence (not modifiable). This involves multi-physician review of planning and recurrence contours, and review of IGRT data (i.e. set-up error, Abbreviations DIR, deformable image registration; IMRT, intensity modulated radiotherapy; adaptive replanning datasets), as well as examination of IRB, institutional review board; pCT, planning CT; rCT, recurrence CT; rGTV, the follow-up interval between surveillance images. By recurrence gross tumor volume; RIR, rigid image registration; ROIs, regions of cataloging type B/D errors, we can then address the rele- interest; TVs, target volumes vant issues dynamically for future patients. For instance, Acknowledgement the only type B patient (i.e. using DIR methodology), This work has been supported by a UICC American Cancer Society was noted on secondary review of diagnostic imaging to Beginning Investigators Fellowship funded by the American Cancer Society. have subsequent intracranial extension, route of failure, Funding sources and financial disclosures despite optimum delineation and dose coverage. Dr. Fuller receives grant support from the National Institutes of Health The secondary qualitative review by expert radiation on- Cancer Center Support (Core) Grant CA016672 to The University of Texas MD cologists (CDF, DIR) of all the clinical and imaging data of Anderson Cancer Center, the National Institutes of Health/National Cancer Institute’s Paul Calabresi Clinical Oncology Award Program (K12 CA088084- the four additional recurrences that were classified as per- 06) and Clinician Scientist Loan Repayment Program (L30 CA136381-02); the ipheral high dose (type B) using RIR while were type A SWOG/Hope Foundation Dr. Charles A. Coltman, Jr., Fellowship in Clinical using DIR, concurred with DIR classification that Trials; a General Electric Healthcare/MD Anderson Center for Advanced Mohamed et al. Radiation Oncology (2016) 11:95 Page 9 of 10 Biomedical Imaging In-Kind Award; an Elekta AB/MD Anderson Department 4. Schoenfeld GO, Amdur RJ, Morris CG, Li JG, Hinerman RW, Mendenhall WM. of Radiation Oncology Seed Grant: The Center for Radiation Oncology Patterns of failure and toxicity after intensity-modulated radiotherapy for Research at MD Anderson Cancer Center; and the MD Anderson Institutional head and neck cancer. Int J Radiat Oncol Biol Phys. 2008;71:377–85. Research Grant Program. 5. Hong TS, Tome WA, Harari PM. Heterogeneity in head and neck IMRT target design and clinical practice. Radiother Oncol. 2012;103:92–8. Availability of data and material 6. Shakam A, Scrimger R, Liu D, Mohamed M, Parliament M, Field GC, et al. The authors confirm availability of anonymized images of failure and Dose-volume analysis of locoregional recurrences in head and neck IMRT, as planning CTs upon request. determined by deformable registration: a prospective multi-institutional trial. Radiother Oncol. 2011;99:101–7. 7. Due AK, Vogelius IR, Aznar MC, Bentzen SM, Berthelsen AK, Korreman Authors’ contributions SS, et al. Methods for estimating the site of origin of locoregional All listed co-authors performed the following: 1. Substantial contributions to the recurrence in head and neck squamous cell carcinoma. Strahlenther conception or design of the work; or the acquisition, analysis, or interpretation Onkol. 2012;188:671–6. of data for the work; 2. Drafting the work or revising it critically for important 8. Raktoe SA, Dehnad H, Raaijmakers CP, Braunius W, Terhaard CH. Origin of intellectual content; 3. Final approval of the version to be published; 4. tumor recurrence after intensity modulated radiation therapy for oropharyngeal Agreement to be accountable for all aspects of the work in ensuring that squamous cell carcinoma. Int J Radiat Oncol Biol Phys. 2013;85:136–41. questions related to the accuracy or integrity of any part of the work are 9. Dawson LA, Anzai Y, Marsh L, Martel MK, Paulino A, Ship JA, et al. Patterns appropriately investigated and resolved. Specific additional individual cooperative of local-regional recurrence following parotid-sparing conformal and effort contributions to study/manuscript design/execution/interpretation, in segmental intensity-modulated radiotherapy for head and neck cancer. Int J addition to all criteria above are listed as follows: ASRM- Undertook clinical and Radiat Oncol. 2000;46:1117–26. imaging data collection; direct oversight of all image registration/segmentation, and data collection workflow, drafted initial manuscript, and participated in data 10. Chao KSC, Ozyigit G, Tran BN, Cengiz M, Dempsey JF, Low DA. Patterns of analysis and interpretation of data. DIR- Co-primary investigator; with CDF failure in patients receiving definitive and postoperative IMRT for head-and- conceived project and interpreted study results, direct and final oversight of neck cancer. Int J Radiat Oncol. 2003;55:312–21. imaging and clinical data collection. GBG, EKU, ASG, MJA, AMB, BMB, AGE, JP- 11. Yao M, Dornfeld KJ, Buatti JM, Skwarchuk M, Tan H, Nguyen T, et al. Direct patient care provision, direct imaging assessment and clinical data Intensity-modulated radiation treatment for head-and-neck squamous cell collection; interpretation and analytic support. CDF- Corresponding author; carcinoma–the University of Iowa experience. Int J Radiat Oncol Biol Phys. co-primary investigator; conceived, coordinated, and directed all study activities, 2005;63:410–21. responsible for data collection, project integrity, manuscript content and 12. Daly ME, Lieskovsky Y, Pawlicki T, Yau J, Pinto H, Kaplan M, et al. Evaluation editorial oversight and correspondence. of patterns of failure and subjective salivary function in patients treated with intensity modulated radiotherapy for head and neck squamous cell carcinoma. Head Neck J Sci Spec. 2007;29:211–20. Competing interests These funding entities played no role in designing the study; collecting, 13. Sanguineti G, Gunn GB, Endres EJ, Chaljub G, Cheruvu P, Parker B. Patterns analyzing, or interpreting its data; writing the manuscript; or making the of locoregional failure after exclusive IMRT for oropharyngeal carcinoma. Int decision to submit the report for publication. J Radiat Oncol Biol Phys. 2008;72:737–46. 14. Crum WR, Hartkens T, Hill DL. Non-rigid image registration: theory and practice. Br J Radiol. 2004;77(Spec No 2):S140–53. Consent for publication 15. Castadot P, Lee JA, Parraga A, Geets X, Macq B, Gregoire V. Comparison Not applicable of 12 deformable registration strategies in adaptive radiation therapy for the treatment of head and neck tumors. Radiother Oncol. 2008;89: Ethics approval and consent to participate 1–12. This study was initiated after the approval of the institutional review board at 16. Lee C, Langen KM, Lu W, Haimerl J, Schnarr E, Ruchala KJ, et al. Evaluation The University of Texas, MD Anderson Cancer Center of the study protocol of geometric changes of parotid glands during head and neck cancer #PA12-0168. radiotherapy using daily MVCT and automatic deformable registration. Radiother Oncol. 2008;89:81–8. Author details 17. Huger S, Graff P, Harter V, Marchesi V, Royer P, Diaz JC, et al. Evaluation of Head and Neck Section, Division of Radiation Oncology, Department of the Block Matching deformable registration algorithm in the field of head- Radiation Oncology, The University of Texas M.D. Anderson Cancer Center, and-neck adaptive radiotherapy. Phys Med. 2014;30(3):301–8. Box 00971515 Holcombe Blvd, Houston, TX 77030, USA. Department of 18. Mohamed AS, Ruangskul MN, Awan MJ, Baron CA, Kalpathy-Cramer J, Clinical Oncology and nuclear medicine, Faculty of Medicine, Alexandria Castillo R, et al. Quality Assurance Assessment of Diagnostic and Radiation University, Alexandria, Egypt. Department of Radiation Oncology, Case Therapy-Simulation CT Image Registration for Head and Neck Radiation Western Reserve University, Cleveland, OH, USA. Department of Radiation Therapy: Anatomic Region of Interest-based Comparison of Rigid and Oncology, Beykent University, Istanbul, Turkey. Graduate School of Deformable Algorithms. Radiology. 2015;274(3):752–63. Biomedical Science, University of Texas Health Science Center, Houston, TX, 19. Johnson B, Sales L, Winston A, Liao J, Laramore G, Parvathaneni U. USA. Fabrication of customized tongue-displacing stents: considerations for use in patients receiving head and neck radiotherapy. J Am Dent Assoc. Received: 15 May 2016 Accepted: 22 July 2016 2013;144:594–600. 20. Goel A, Tripathi A, Chand P, Singh SV, Pant MC, Nagar A. Use of positioning stents in lingual carcinoma patients subjected to radiotherapy. Int J References Prosthodont. 2010;23:450–2. 1. Gregoire V, Coche E, Cosnard G, Hamoir M, Reychler H. Selection and 21. Tao R, Fuller CD, Gunn GB, Beadle BM, Phan J, Frank SJ, et al. Real-time peer delineation of lymph node target volumes in head and neck conformal review quality assurance conferences incorporating physical examination for radiotherapy. Proposal for standardizing terminology and procedure based head-and-neck cancer radiation therapy result in clinically meaningful target on the surgical experience. Radiother Oncol. 2000;56:135–50. volume alteration: results of a prospective volumetric analysis. Int J Radiat 2. Eisbruch A, Marsh LH, Dawson LA, Bradford CR, Teknos TN, Chepeha DB, Oncol Biol Phys. 2012;84:S151. et al. Recurrences near base of skull after IMRT for head-and-neck cancer: 22. Rosenthal DI, Asper JA, Barker JL, Garden AS, Chao KSC, Morrison WH, et al. implications for target delineation in high neck and for parotid gland Importance of patient examination to clinical quality assurance in head and sparing. Int J Radiat Oncol Biol Phys. 2004;59:28–42. neck radiation oncology. Head Neck J Sci Spec. 2006;28:967–73. 3. Suzuki M, Nishimura Y, Nakamatsu K, Okumura M, Hashiba H, Koike R, et al. Analysis of interfractional set-up errors and intrafractional organ motions 23. International Commission on Radiation Units and Measurements. ICRU during IMRT for head and neck tumors to define an appropriate planning Report 62. Prescribing, recording, and reporting photon beam therapy target volume (PTV)- and planning organs at risk volume (PRV)-margins. (Supplement to ICRU Report 50)ICRU. Bethesda: Oxford University Press, Radiother Oncol. 2006;78:283–90. Oxford, United Kingdom; 1999. Mohamed et al. Radiation Oncology (2016) 11:95 Page 10 of 10 24. International Commission on Radiation Units and Measurements. ICRU Report 50. Prescribing, recording, and reporting photon beam therapy ICRU. Bethesda: Oxford University Press, Oxford, United Kingdom; 1993. 25. International Commission on Radiation Units and Measurements. ICRU Report 83: prescribing, recording and reporting photon-beam intensity- modulated radiation therapy (IMRT). J ICRU. 2010;10:1–106. 26. Espana S, Paganetti H. Uncertainties in planned dose due to the limited voxel size of the planning CT when treating lung tumors with proton therapy. Phys Med Biol. 2011;56:3843–56. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Radiation Oncology Springer Journals

Methodology for analysis and reporting patterns of failure in the Era of IMRT: head and neck cancer applications

Download PDF
10 pages

Loading next page...
 
/lp/springer-journals/methodology-for-analysis-and-reporting-patterns-of-failure-in-the-era-NuUVfDDkKN

References (26)

Publisher
Springer Journals
Copyright
Copyright © 2016 by The Author(s).
Subject
Biomedicine; Cancer Research; Oncology; Radiotherapy
eISSN
1748-717X
DOI
10.1186/s13014-016-0678-7
pmid
27460585
Publisher site
See Article on Publisher Site

Abstract

Background: The aim of this study is to develop a methodology to standardize the analysis and reporting of the patterns of loco-regional failure after IMRT of head and neck cancer. Material and Methods: Twenty-one patients with evidence of local and/or regional failure following IMRT for head-and-neck cancer were retrospectively reviewed under approved IRB protocol. Manually delineated recurrent gross disease (rGTV) on the diagnostic CT documenting recurrence (rCT) was co-registered with the original planning CT (pCT) using both deformable (DIR) and rigid (RIR) image registration software. Subsequently, mapped rGTVs were compared relative to original planning target volumes (TVs) and dose using a centroid-based approaches. Failures were then classified into five types based on combined spatial and dosimetric criteria; A (central high dose), B (peripheral high dose), C (central elective dose), D (peripheral elective dose), and E (extraneous dose). Results: A total of 26 recurrences were identified. Using DIR, recurrences were assigned to more central TVs compared to RIR as detected using the spatial centroid-based method (p = 0.0002). rGTVs mapped using DIR had statistically significant higher mean doses when compared to rGTVs mapped rigidly (mean dose 70 vs. 69 Gy, p =0. 03). According to the proposed classification 22 out of 26 failures were of type A (central high dose) as assessed by DIR method compared to 18 out of 26 for the RIR because of the tendencey of RIR to assign failures more peripherally. Conclusions: RIR tends to assigns failures more peripherally. DIR-based methods showed that the vast majority of failures originated in the high dose target volumes and received full prescribed doses suggesting biological rather than technology-related causes of failure. Validated DIR-based registration is recommended for accurate failure characterization and a novel typology-indicative taxonomy is recommended for failure reporting in the IMRT era. Keywords: Patterns of failure, IMRT, Head-and-neck cancer, Deformable image registration Introduction (defined as tumor persistence or recurrence) including in- Intensity-modulated radiation therapy (IMRT) is one of adequate definition of the tumor extension and clinically the most important innovations in modern radiation ther- important target volumes (TVs), uncertainties related to apy and represents a paradigm shift in the treatment of daily positioning, weight loss or deformation of tumor and head and neck cancers (HNCs). However, there are certain normal tissues during the course of treatment, and uncer- hazards that may increase the risk of loco-regional failure tainties in plan optimization, dose calculation and treat- ment delivery [1–5]. The accurate and specific definition of the exact site of * Correspondence: asmohamed@mdanderson.org; cdfuller@mdanderson.org failure, in addition to the radiation dose given to this site Head and Neck Section, Division of Radiation Oncology, Department of is, therefore, mandatory to identify the possible cause(s) Radiation Oncology, The University of Texas M.D. Anderson Cancer Center, Box 00971515 Holcombe Blvd, Houston, TX 77030, USA of failure. The classic definition of failures as “local”,or Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Mohamed et al. Radiation Oncology (2016) 11:95 Page 2 of 10 “regional”, was appropriate in the setting of conventional documented recurrence. A total of 21 cases were ran- radiotherapy using large homogeneous dose-volumes, domly selected from the recurrence dataset based on the but is no longer helpful nor descriptive of distinct types following eligibility criteria: IMRT given for curative in- of failure in patients treated with IMRT [6–8]. tent; treatment of intact tumor (i.e. post-operative cases Several previous efforts have addressed the importance were excluded); equal distribution of various head and of studying the patterns of failure after IMRT treatment of neck subsites (i.e. nasopharynx, oropharynx, hypopharynx, HNCs, [2, 4, 9–13] with most reporting failures as and lateral neck “i.e. neck nodes of unknown primary “infield”, “marginal” or “outfield” based on the percentage site”); radiological evidence of local and/or regional failure; of overlap between the failure volume and the respective available CT scan of failure site prior to any salvage ther- TV on the treatment planning CT (pCT) [4, 9, 10, 12, 13]. apy; and pathologic and/or radiologic evidence of recur- The ability to accurately describe the relation of failure rence (i.e. biopsy, or high SUV on PET). to original TVs and dose mandates a fairly precise method to co-register the diagnostic CT documenting IMRT treatment planning and delivery recurrence (rCT) to the original pCT. However, the ma- All patients had been positioned supine in an individualized jority of the previous studies implemented mainly rigid thermoplastic head and shoulder mask for CT simulation image registration techniques (RIR) [2, 4, 10, 12, 13]. and treatment and a custom dental stent used as an RIR is simple, quick and widely used but it allows only intraoral immobilization device [19, 20]. A treatment for 6° of freedom and doesn’t account for changes in the pCT scan was used for defining TVs. Target volume shapes or relative positions of different regions-of- definition was done in Pinnacle treatment planning interests (ROIs) [14]. Emerging data demonstrate the system (Pinnacle, Phillips Medical Systems, Andover, superiority of deformable image registration (DIR) com- MA), with rigorous multi-physician target delineation pared to RIR in registering pCT to on-treatment CT or and quality assurance [21, 22]. conebeam CT in the setting of image guided radiother- Treatment was uniformly delivered using Varian (Varian apy (IGRT) for HNCs [15–17]. However, very few stud- Medical Systems, Palo Alto, CA) linear accelerators ies addressed DIR software implementation for the delivering 6-MV photons. Three clinical target volumes purpose of registering the diagnostic rCT to the original (CTVs) were typically defined. CTV definitions and dose pCT [6, 7]. prescriptions are summarized in Additional file 1: Our group has recently validated different registration Table S1. Treatment was delivered in a conventional frac- techniques used for co-registering diagnostic contrast en- tionation scheme (average 33 fractions). Patients were hanced head and neck CT to non-contrast planning CT treated using a monoisocentric technique with an antero- and showed DIR was superior for this application [18]. As posterior low-neck supraclavicular field matched to the a continuation of these efforts and to validate DIR as a IMRT fields or using whole neck IMRT for cases where tool to improve accurate definition of the patterns of loco- gross nodes are located at the match line. regional failure in the era of IMRT for HNCs, we sought to undergo the following specific aims: Post-treatment follow up Initial post-treatment evaluations were made at 8–12 1) Develop a workflow methodology to standardize the weeks after therapy completion and subsequently every analysis of HNCs patterns of failure using both 2–3 months for the first year, every 3–4 months for the geometric and dosimetric parameters. second year, and at least twice a year up to 5 years. 2) Assess the impact of registration (rigid vs. deformable) techniques on patterns of failure Loco-regional failure quantitative analytic parameters. Cases where local and/or regional recurrence was re- 3) Develop a granular classification and nomenclature corded had their immediate post-failure diagnostic im- to optimize the accurate reporting of distinct failure ages exported as DICOM files from the clinical PACS typology. system to the treatment planning system, where radio- logical evident recurrent gross disease (rGTV) was Material and methods manually contoured by a radiation oncologist (ASRM) Patients and reviewed by a head and neck service-specific attend- Tumor registry data for patients diagnosed with head ing radiation oncologist (CDF). and neck squamous cell carcinoma, whom were treated by IMRT at The University of Texas, MD Anderson Image registration Cancer Center between 2006 and 2009, were retrospect- For each patient, the rCT or rPET-CT was co-registered ively reviewed under an institutional review board ap- with pCT using both rigid and deformable image registra- proval. 600 patients were identified, of those 103 had a tion techniques. DIR was performed using a commercial Mohamed et al. Radiation Oncology (2016) 11:95 Page 3 of 10 software (ADMIRE, Elekta AB, 2015) validated previously the Pinnacle planning system to rigidly align rCT to by our group for the registration of contrast-enhanced pCT, following that rGTVs where exported to patient’s diagnostic CT to non-contrast enhanced planning CT plan where dose volume histograms (DVHs) and rGTV [18]. Deformation vector fields were obtained from DIR centroids were generated and analysis metrics were cal- algorithm, mapping the deformation of the rCT onto the culated. Figure 1 illustrates the workflow process of our pCT. Subsequently, in a custom written Matlab routine registration methodology. (MATLAB R2013a, The MathWorks Inc., Natick, MA, 2013), pCT; dose grid; original plan TVs; rCT; and rGTVs Analysis of failure metrics were imported. The deformation fields were then applied For both RIR and DIR mapped rGTVs the following to rGTVs segmented on the rCT to convert them into ‘de- metrics were evaluated: formed rGTVs’ on the pCT. 1) Recurrence volume, 2) Location of the centroid Evaluation of deformed rGTVs relative to original relative to planning TV: Centroid is the central voxel of planning TVs was done using both centroid-based the recurrence volume plus added 2 mm margin to ac- method that assumed the center of mass of rGTV was count for registration error, 3) Spatial relationship of the origin of the recurrence volume and its location was rGTV centroids to IMRT/supraclavicular match line and compared relative to planning TV after applying deform- ipsilateral parotid in case of peri-parotid failure, and 4) ation vector fields (DVF). Simultaneously, RIR was per- Mean and maximum dose to rGTVs, dose to 95 % fail- formed using the rigid co-registration tool available in ure volume (fD95%), and mean dose to centroid volume. Fig. 1 Workflow process of patterns of loco-regional failure registration process Mohamed et al. Radiation Oncology (2016) 11:95 Page 4 of 10 Classification of failure  Type D: Peripheral elective dose failure, where the In order to refine our reporting and quality assurance failure centroid originates from elective dose TV and practices using a standard nomenclature, we developed a its fD95% receives <95 % dose prescribed to the granular typology of failure categories relative to the respective TV. planning TV and dose. As illustrated in Fig. 2, failures  Type E: Extraneous dose failure, where the failure were classified into five types based on combined spatial centroid originates outside all TVs. and dosimetric criteria: For patients treated with low-neck supraclavicular field Type A: Central high dose failure, where the mapped matched to the IMRT fields, two additional types were failure centroid originates in high dose TV and the added: dose to 95 % failure volume (fD95%) is ≥95 % dose prescribed to corresponding TV of origin.  Type F: Junctional failures at the site of IMRT/ Type B: Peripheral high dose failure, where the supraclavicular match line. mapped failure centroid originates from high dose  Type G: Low neck failures at the site of low-neck TV but its fD95% receives <95 % dose prescribed to supraclavicular field. this TV. Type C: Central elective dose failure, where the Statistical analysis failure centroid originates from elective dose TV Non parametric statistics were used to compare analysis and its fD95% receives ≥95 % dose prescribed to the metrics for centroid locations and dosimetric parameters respective TV. of failures mapped using RIR versus DIR registration Fig. 2 Classification scheme of IMRT patterns of failure using combined centroid based geometric method coupled with the dosimetric parameters. Panel a) shows an example of types A (central high dose) and C (central elective dose) failures, panel b) shows an example of types B (peripheral high dose) and D (peripheral elective dose) failures, and panel c) shows an example of type E (extraneous dose) failure. Mohamed et al. Radiation Oncology (2016) 11:95 Page 5 of 10 techniques. A p-value ≤ 0.05 was deemed significant. Table 1 Patient demographics, disease, and treatment characteristics Statistical assessment and data tabulation was performed using JMP v 11Pro (SAS institute, Cary, NC). Total n = 21 (%) Results Age (years) Patients Median 58 A total of 21 patients with HNSCC were included in this Range 30–75 pilot methodology/workflow development study. Median Time to Failure (months) age was 58 years (range 30–75), and 86 % were men. Median 12 Patient, disease, and treatment characteristics are pre- sented in Table 1. Recurrences were delineated using Range 5–69 diagnostic contrast-enhanced CT in 16 patients and Sex using PET-CT in 5 patients. Male 18 (86) Female 3 (14) Spatial/dosimetric failure mapping Origin Spatial mapping Nasopharynx 6 (28) A total number of 26 rGTVs were delineated. Mean rGTVs volume was 12.5 cm (range 1–105). The regis- Oropharynx 5 (24) tration method independently affected the spatial loca- Hypopharynx 5 (24) tion of mapped failures. Failures mapped using DIR Unknown primary 5 (24) were significantly assigned to more central TVs com- T-category pared to failures mapped using RIR. 38 % of centroids T0 5 (24) (n = 10) mapped using RIR were located peripheral to T1 1 (5) the same centroids mapped using DIR (p = 0.0002). Table 2 illustrates the sites and geometric details of all T2 7 (33) failures mapped to the pCT. T3 5 (24) T4 3 (14) Dosimetric mapping N-category rGTVs mapped using DIR had statistically significant N0 1 (5) higher mean doses when compared to rGTVs mapped N1 5 (24) rigidly (mean dose 70 vs. 69 Gy, p = 0.03) while compari- son of mean fD95% was not statistically significant N2 12 (57) (mean fD95% 68 vs. 66 Gy, p = 0.07), and comparison of N3 3 (14) maximum, and centroid doses showed no-significant dif- Treatment ferences between both registration methods (p = 0.7 and Radiation alone 4 (19) 0.4, respectively). Additional file 1: Table S2 shows the Concurrent ChemoRadiation 9 (43) dosimetric details of all failures. Induction Chemotherapy + Radiation 1 (5) Classification of failure Induction Chemotherapy + Concurrent ChemoRadiation 7 (33) Based on the proposed classification of failure using Radiation dose both the spatial location of the centroids of the mapped Mean (SD) 69.2 (1.7) failure volumes coupled with the dosimetric parameters Radiation fractions (as illustrated in Fig. 2), 22 (84.6 %) out of the 26 failures Mean (SD) 33 (2) mapped using DIR were of type A, one of type B, 2 of type C, and one of type G. Whereas, 18 (69 %) out of the 26 failure mapped using RIR were of type A, 5 of located peripheral to the same centroids mapped using type B, 2 of type C and one of type G. Figure 3 illustrates DIR as shown in Table 2. However, after adding the the difference in classification using both registration dosimetric component of analysis, only 4 of those 10 methods. There was no type F (junctional) failures in pa- RIR mapped rGTVs were peripheral high dose failures tient subset treated using anteroposterior low-neck (type B) and the other 6 were central high dose failure supraclavicular field matched to the IMRT fields. Addition- (type A) because despite the centroids were spatially per- ally, no peri-parotid failures were detected. ipheral in location to the respective DIR ones but dosi- This combined spatial/dosimetric analysis shows that metrically, the rGTVs 95 % volumes still had ≥95 % while 10 centroids (38 %) of RIR mapped rGTVs were dose. Figure 4 shows an example of the differences in Mohamed et al. Radiation Oncology (2016) 11:95 Page 6 of 10 Table 2 Geometric details of failed rGTVs reference. However, such a reporting language gives no information regarding radiation fields/volumes, or deliv- n. Percent ered dose. In the pre-IMRT era, when large fields of N. of recurrences 26 homogenous dose were used, the definition of “in-field” Recurrence volume failure (i.e. within the field borders) or “marginal” failure Mean (SD) 12.5 (23) (i.e. adjacent to the block edges) were intuitive descrip- Location of centroid using RIR tors relating treatment parameters to sites of failure. GTV 12 (46) However, in the current era of conformal therapy [23, 24], CTV1 7 (27) dose gradients, and multiple TVs make relating spatially accurate information about dose and recurrence far more CTV2 1 (4) complicated for IMRT plans. In the same way that a stan- CTV3 1 (4) dardized method for analysis and reporting for TVs has PTV1 4 (15) been undertaken successfully [23–25], a similar effort is Supraclavicular field 1 (4) desirable for pattern of failure reporting. In our opinion, Location of centroid using DIR reporting failure using only anatomic/field referents is in- GTV 22 (84) sufficient for complex multi-volume/multi-dose plans, and obscures clinically useful information which might CTV1 1 (4) lead to improvements in future studies. CTV2 1 (4) Likewise, the requirement of rigorous assurance for CTV3 1 (4) correctly localizing disease after-therapy is increased in Supraclavicular field 1 (4) terms of required spatial accuracy. The steep dose gradi- Abbreviations: DIR Deformable image registration, RR Rigid Registration, GTV ents of modern IMRT plans and proximate transition gross tumor volume, CTV clinical target volume, PTV planning target volume from high-risk CTVs to intermediate- or lower-risk CTVs implies inaccurate registration will erroneously as- spatial and dosimetric parameters for a DIR versus RIR sign the location of failure to incorrect dose/prescription mapped failure. Those 4 rGTVs were seen in the follow- volume. We, in fact, show RIR for IMRT plans resulted ing patients: two nasopharyngeal (one primary “Fig. 4” in incorrect localization relative to prior TVs and dose and one nodal site); one oropharyngeal (primary site); in 16 % of failures in the current pilot dataset. Conse- and one unkown primary (nodal site). The secondary quently, this study presents a methodology and workflow qualitative review by expert radiation oncologists (CDF, that involves the application of quality assured DIR soft- DIR) of those 4 patients agreed with DIR classification ware as a tool to standardize co-registration and to cor- that those recurrences are actually central rather than rectly attribute sites of loco-regional failure. peripheral in origin. Almost all previous studies have used RIR to de- scribe the patterns of loco-regional failure after IMRT Discussion [2, 10, 12, 13]. Chao et al. [10] reported 17/126 loco- Traditionally, failure reporting for HNCs has simply regional failures treated by definitive or postoperative classified disease as “local”, “regional”,or “locoregional”, IMRT; 53 % of failures were inside CTV1, 12 % mar- thus relating location of failure to a crude anatomic ginal to CTV1, 6 % marginal to CTV2, and 28 % Fig. 3 Bar chart illustrating the difference in failure classification using rigid (RIR) vs. deformable (DIR) image registration methods Mohamed et al. Radiation Oncology (2016) 11:95 Page 7 of 10 Fig. 4 A case of T2N0 Nasopharyngeal carcinoma recurred 63 months after IMRT. The upper panel shows the axial, coronal and sagittal images of a RIR mapped rGTV on the original pCT where its centroid is located at CTV1and the 95 % rGTV volume contained on more peripheral PTV2 (contour not shown). The middle panel shows DIR mapped rGTV on the original pCT where its centroid located at GTV and the 95 % rGTV volume contained on more peripheral CTV2. The lower panel shows RIR and DIR mapped rGTVs overlaid to plan isodose line. Note that RIR rGTV fD95% extends beyond the 95 % isodose line “66.5 Gy” (red arrow in sagittal image) which would erroneously characterize it as type B failure, while in fact DIR shows it as a type A failure (i.e. the fD95% of DIR mapped rGTV is completely encapsulated with 95 % isodose line, shown by white arrow in sagittal image) were out of field. Eisbruch et al. [2] reported on 21 RIR. Our previous work [18], as well as the current recurrences in 133 patients with non-nasopharyngeal study, confirm the qualitative superiority, in HNC ap- HNCs treated with parotid-sparing IMRT; 17/21 were plications, of DIR for CT-CT registration. in-field. Daly et al. [12] reported on 69 HNCs treated In our classification scheme, we designed a combined with parotid-sparing IMRT; 8 patients developed a geometric/dosimetric typology definition to avoid the loco-regional failure, 7 relapsed within the high-dose drawbacks of using each method separately. Centroid CTV, with one junctional failure observed. Sanguineti only method suppose a single point of origin and ignore et al. [13] described the patterns of failure for 50 pa- the dose given to the whole area of recurrence while the tients with IMRT for oropharyngeal SCC; 5 recur- dosimetric only analysis is agnostic to the geometric re- rences were related to high dose regions while 4 were currence origin. Due et al. [7] previously reported that at the low dose regions. All these reports relied on focal methods, such as the centroid method we used, are RIR, known to be less spatially accurate than DIR more accurate to localize the origins of loco-regional re- [18]; it is conceivable these results might be altered if currences than volume overlap methods, which may in- DIR methods were used. Due et al. [7] reported that correctly assign recurrences to more peripheral TVs. DIR showed slightly better reproducibility in identifi- Raktoe et al. [8] further confirmed the superiority of cation of the site of recurrence origin compared to focal methods like centroid expansion to the volumetric Mohamed et al. Radiation Oncology (2016) 11:95 Page 8 of 10 method in identifying the origin of loco-regional recur- those recurrences are actually central rather than per- rences. The combined method we used identify the esti- ipheral in origin. mated site of recurrence origin relative to the respective In this limited pilot dataset, our results showed the TV in the planning CT and then compare the dose to majority (84.6 %) of DIR mapped failures were of type A the mapped recurrence volume with the dose prescribed indicating, that biological, non-technically/non-operator to the TV of origin. Using this method, our results dependant explanations for failure predominated. How- showed that DIR significantly assigned failures to more ever, using RIR type A failures would have been errone- central TVs and doses compared to RIR concordant with ously reported as comprising only 69 %. These results the results of Due et al. [7]. assert the need for a robust, quality assured image regis- Our proposed nomenclature allows granular reporting tration technique, as error in the registration process of different types of failure. In our classification, type A would invalidate subsequent results and thus might “central high dose” failures, are considered to be bio- deceptively indicate a greater rate of technical/operator- logical failures, as they likely represent resistance to attributable therapy failures than DIR demonstrates. The maximal therapy, and thus could not conceivably be pre- current study, while underpowered to make clinical ex- vented by technical/operator dependant processes in- trapolations due to limited number of patients, nonethe- cluding IMRT QA or delineation alteration. Type A less serves as a benchmark to describe our standardized failures motivate future investigation of alternative treat- analytic and reporting method. Already, RTOG 1216, for ment stratgies (e.g. integration of novel targeted drug example, contains provisions regarding collection of im- therapies or dose escalation to identifiable biologically aging data post-failure [26], which will allow careful ana- aggressive subvolumes). Likewise, type E “extraneous lysis, and process quality improvement for future trials dose” failures cannot be modified by IMRT QA pro- and large scale datasets. cesses. They represent a possible diagnostic or decision error rather than a target delineation error (i.e. “One will Conclusions never hit what one does not aim at.”). However, type B, Rigid image registration method tends to assigns failures C, and D failures are of a special concern since they more peripherally compared with deformable method. entail potential technical or radiotherapy process fail- Using DIR, the vast majority of failures in the presented ures. Type C “central elective (intermediate or low) pilot study originated in the high dose target volumes dose” failures may be prevented by prescribing higher and received full prescribed doses suggesting biological doses (i.e. shifting to higher CTV levels). Importantly, rather than technology-related causes of failure. We type B “peripheral high dose” or D “peripheral elective heavily recommend a validated DIR-based registration dose” failures necessitate a rigorous QA process includ- technique in addition to granular combined geometric- ing triple DIR registration of pre-therapy diagnostic im- and dosimetric-based failure characterization using aging (diagnostic CT, MRI, and/or PET-CT) to pCT and novel typology-indicative taxonomy as a standard part of the earliest rCT, to assess the potential causes: potential large-scale patterns of failure reporting in the IMRT era. target delineation or dose delivery error (modifiable) ver- sus overgrown recurrence that represents actual type A Additional file or C failure which is converted to type B or D, respect- ively, due to rapidly progressive disease or neglected late Additional file 1: Table S1. IMRT target volume definitions and dose prescription. Table S2. Dosimetric patterns of failure. (DOCX 16 kb) diagnosed recurrence (not modifiable). This involves multi-physician review of planning and recurrence contours, and review of IGRT data (i.e. set-up error, Abbreviations DIR, deformable image registration; IMRT, intensity modulated radiotherapy; adaptive replanning datasets), as well as examination of IRB, institutional review board; pCT, planning CT; rCT, recurrence CT; rGTV, the follow-up interval between surveillance images. By recurrence gross tumor volume; RIR, rigid image registration; ROIs, regions of cataloging type B/D errors, we can then address the rele- interest; TVs, target volumes vant issues dynamically for future patients. For instance, Acknowledgement the only type B patient (i.e. using DIR methodology), This work has been supported by a UICC American Cancer Society was noted on secondary review of diagnostic imaging to Beginning Investigators Fellowship funded by the American Cancer Society. have subsequent intracranial extension, route of failure, Funding sources and financial disclosures despite optimum delineation and dose coverage. Dr. Fuller receives grant support from the National Institutes of Health The secondary qualitative review by expert radiation on- Cancer Center Support (Core) Grant CA016672 to The University of Texas MD cologists (CDF, DIR) of all the clinical and imaging data of Anderson Cancer Center, the National Institutes of Health/National Cancer Institute’s Paul Calabresi Clinical Oncology Award Program (K12 CA088084- the four additional recurrences that were classified as per- 06) and Clinician Scientist Loan Repayment Program (L30 CA136381-02); the ipheral high dose (type B) using RIR while were type A SWOG/Hope Foundation Dr. Charles A. Coltman, Jr., Fellowship in Clinical using DIR, concurred with DIR classification that Trials; a General Electric Healthcare/MD Anderson Center for Advanced Mohamed et al. Radiation Oncology (2016) 11:95 Page 9 of 10 Biomedical Imaging In-Kind Award; an Elekta AB/MD Anderson Department 4. Schoenfeld GO, Amdur RJ, Morris CG, Li JG, Hinerman RW, Mendenhall WM. of Radiation Oncology Seed Grant: The Center for Radiation Oncology Patterns of failure and toxicity after intensity-modulated radiotherapy for Research at MD Anderson Cancer Center; and the MD Anderson Institutional head and neck cancer. Int J Radiat Oncol Biol Phys. 2008;71:377–85. Research Grant Program. 5. Hong TS, Tome WA, Harari PM. Heterogeneity in head and neck IMRT target design and clinical practice. Radiother Oncol. 2012;103:92–8. Availability of data and material 6. Shakam A, Scrimger R, Liu D, Mohamed M, Parliament M, Field GC, et al. The authors confirm availability of anonymized images of failure and Dose-volume analysis of locoregional recurrences in head and neck IMRT, as planning CTs upon request. determined by deformable registration: a prospective multi-institutional trial. Radiother Oncol. 2011;99:101–7. 7. Due AK, Vogelius IR, Aznar MC, Bentzen SM, Berthelsen AK, Korreman Authors’ contributions SS, et al. Methods for estimating the site of origin of locoregional All listed co-authors performed the following: 1. Substantial contributions to the recurrence in head and neck squamous cell carcinoma. Strahlenther conception or design of the work; or the acquisition, analysis, or interpretation Onkol. 2012;188:671–6. of data for the work; 2. Drafting the work or revising it critically for important 8. Raktoe SA, Dehnad H, Raaijmakers CP, Braunius W, Terhaard CH. Origin of intellectual content; 3. Final approval of the version to be published; 4. tumor recurrence after intensity modulated radiation therapy for oropharyngeal Agreement to be accountable for all aspects of the work in ensuring that squamous cell carcinoma. Int J Radiat Oncol Biol Phys. 2013;85:136–41. questions related to the accuracy or integrity of any part of the work are 9. Dawson LA, Anzai Y, Marsh L, Martel MK, Paulino A, Ship JA, et al. Patterns appropriately investigated and resolved. Specific additional individual cooperative of local-regional recurrence following parotid-sparing conformal and effort contributions to study/manuscript design/execution/interpretation, in segmental intensity-modulated radiotherapy for head and neck cancer. Int J addition to all criteria above are listed as follows: ASRM- Undertook clinical and Radiat Oncol. 2000;46:1117–26. imaging data collection; direct oversight of all image registration/segmentation, and data collection workflow, drafted initial manuscript, and participated in data 10. Chao KSC, Ozyigit G, Tran BN, Cengiz M, Dempsey JF, Low DA. Patterns of analysis and interpretation of data. DIR- Co-primary investigator; with CDF failure in patients receiving definitive and postoperative IMRT for head-and- conceived project and interpreted study results, direct and final oversight of neck cancer. Int J Radiat Oncol. 2003;55:312–21. imaging and clinical data collection. GBG, EKU, ASG, MJA, AMB, BMB, AGE, JP- 11. Yao M, Dornfeld KJ, Buatti JM, Skwarchuk M, Tan H, Nguyen T, et al. Direct patient care provision, direct imaging assessment and clinical data Intensity-modulated radiation treatment for head-and-neck squamous cell collection; interpretation and analytic support. CDF- Corresponding author; carcinoma–the University of Iowa experience. Int J Radiat Oncol Biol Phys. co-primary investigator; conceived, coordinated, and directed all study activities, 2005;63:410–21. responsible for data collection, project integrity, manuscript content and 12. Daly ME, Lieskovsky Y, Pawlicki T, Yau J, Pinto H, Kaplan M, et al. Evaluation editorial oversight and correspondence. of patterns of failure and subjective salivary function in patients treated with intensity modulated radiotherapy for head and neck squamous cell carcinoma. Head Neck J Sci Spec. 2007;29:211–20. Competing interests These funding entities played no role in designing the study; collecting, 13. Sanguineti G, Gunn GB, Endres EJ, Chaljub G, Cheruvu P, Parker B. Patterns analyzing, or interpreting its data; writing the manuscript; or making the of locoregional failure after exclusive IMRT for oropharyngeal carcinoma. Int decision to submit the report for publication. J Radiat Oncol Biol Phys. 2008;72:737–46. 14. Crum WR, Hartkens T, Hill DL. Non-rigid image registration: theory and practice. Br J Radiol. 2004;77(Spec No 2):S140–53. Consent for publication 15. Castadot P, Lee JA, Parraga A, Geets X, Macq B, Gregoire V. Comparison Not applicable of 12 deformable registration strategies in adaptive radiation therapy for the treatment of head and neck tumors. Radiother Oncol. 2008;89: Ethics approval and consent to participate 1–12. This study was initiated after the approval of the institutional review board at 16. Lee C, Langen KM, Lu W, Haimerl J, Schnarr E, Ruchala KJ, et al. Evaluation The University of Texas, MD Anderson Cancer Center of the study protocol of geometric changes of parotid glands during head and neck cancer #PA12-0168. radiotherapy using daily MVCT and automatic deformable registration. Radiother Oncol. 2008;89:81–8. Author details 17. Huger S, Graff P, Harter V, Marchesi V, Royer P, Diaz JC, et al. Evaluation of Head and Neck Section, Division of Radiation Oncology, Department of the Block Matching deformable registration algorithm in the field of head- Radiation Oncology, The University of Texas M.D. Anderson Cancer Center, and-neck adaptive radiotherapy. Phys Med. 2014;30(3):301–8. Box 00971515 Holcombe Blvd, Houston, TX 77030, USA. Department of 18. Mohamed AS, Ruangskul MN, Awan MJ, Baron CA, Kalpathy-Cramer J, Clinical Oncology and nuclear medicine, Faculty of Medicine, Alexandria Castillo R, et al. Quality Assurance Assessment of Diagnostic and Radiation University, Alexandria, Egypt. Department of Radiation Oncology, Case Therapy-Simulation CT Image Registration for Head and Neck Radiation Western Reserve University, Cleveland, OH, USA. Department of Radiation Therapy: Anatomic Region of Interest-based Comparison of Rigid and Oncology, Beykent University, Istanbul, Turkey. Graduate School of Deformable Algorithms. Radiology. 2015;274(3):752–63. Biomedical Science, University of Texas Health Science Center, Houston, TX, 19. Johnson B, Sales L, Winston A, Liao J, Laramore G, Parvathaneni U. USA. Fabrication of customized tongue-displacing stents: considerations for use in patients receiving head and neck radiotherapy. J Am Dent Assoc. Received: 15 May 2016 Accepted: 22 July 2016 2013;144:594–600. 20. Goel A, Tripathi A, Chand P, Singh SV, Pant MC, Nagar A. Use of positioning stents in lingual carcinoma patients subjected to radiotherapy. Int J References Prosthodont. 2010;23:450–2. 1. Gregoire V, Coche E, Cosnard G, Hamoir M, Reychler H. Selection and 21. Tao R, Fuller CD, Gunn GB, Beadle BM, Phan J, Frank SJ, et al. Real-time peer delineation of lymph node target volumes in head and neck conformal review quality assurance conferences incorporating physical examination for radiotherapy. Proposal for standardizing terminology and procedure based head-and-neck cancer radiation therapy result in clinically meaningful target on the surgical experience. Radiother Oncol. 2000;56:135–50. volume alteration: results of a prospective volumetric analysis. Int J Radiat 2. Eisbruch A, Marsh LH, Dawson LA, Bradford CR, Teknos TN, Chepeha DB, Oncol Biol Phys. 2012;84:S151. et al. Recurrences near base of skull after IMRT for head-and-neck cancer: 22. Rosenthal DI, Asper JA, Barker JL, Garden AS, Chao KSC, Morrison WH, et al. implications for target delineation in high neck and for parotid gland Importance of patient examination to clinical quality assurance in head and sparing. Int J Radiat Oncol Biol Phys. 2004;59:28–42. neck radiation oncology. Head Neck J Sci Spec. 2006;28:967–73. 3. Suzuki M, Nishimura Y, Nakamatsu K, Okumura M, Hashiba H, Koike R, et al. Analysis of interfractional set-up errors and intrafractional organ motions 23. International Commission on Radiation Units and Measurements. ICRU during IMRT for head and neck tumors to define an appropriate planning Report 62. Prescribing, recording, and reporting photon beam therapy target volume (PTV)- and planning organs at risk volume (PRV)-margins. (Supplement to ICRU Report 50)ICRU. Bethesda: Oxford University Press, Radiother Oncol. 2006;78:283–90. Oxford, United Kingdom; 1999. Mohamed et al. Radiation Oncology (2016) 11:95 Page 10 of 10 24. International Commission on Radiation Units and Measurements. ICRU Report 50. Prescribing, recording, and reporting photon beam therapy ICRU. Bethesda: Oxford University Press, Oxford, United Kingdom; 1993. 25. International Commission on Radiation Units and Measurements. ICRU Report 83: prescribing, recording and reporting photon-beam intensity- modulated radiation therapy (IMRT). J ICRU. 2010;10:1–106. 26. Espana S, Paganetti H. Uncertainties in planned dose due to the limited voxel size of the planning CT when treating lung tumors with proton therapy. Phys Med Biol. 2011;56:3843–56. Submit your next manuscript to BioMed Central and we will help you at every step: • We accept pre-submission inquiries � Our selector tool helps you to find the most relevant journal � We provide round the clock customer support � Convenient online submission � Thorough peer review � Inclusion in PubMed and all major indexing services � Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit

Journal

Radiation Oncology – Springer Journals

Published: Jul 26, 2016

There are no references for this article.