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

Learn More →

Tumor volume in subcutaneous mouse xenografts measured by microCT is more accurate and reproducible than determined by 18F-FDG-microPET or external caliper

Tumor volume in subcutaneous mouse xenografts measured by microCT is more accurate and... Background: In animal studies tumor size is used to assess responses to anticancer therapy. Current standard for volumetric measurement of xenografted tumors is by external caliper, a method often affected by error. The aim of the present study was to evaluate if microCT gives more accurate and reproducible measures of tumor size in mice compared with caliper measurements. Furthermore, we evaluated the accuracy of tumor volume determined from F- fluorodeoxyglucose ( F-FDG) PET. Methods: Subcutaneously implanted human breast adenocarcinoma cells in NMRI nude mice served as tumor model. Tumor volume (n = 20) was determined in vivo by external caliper, microCT and F-FDG-PET and subsequently reference volume was determined ex vivo. Intra- observer reproducibility of the microCT and caliper methods were determined by acquiring 10 repeated volume measurements. Volumes of a group of tumors (n = 10) were determined independently by two observers to assess inter-observer variation. Results: Tumor volume measured by microCT, PET and caliper all correlated with reference volume. No significant bias of microCT measurements compared with the reference was found, whereas both PET and caliper had systematic bias compared to reference volume. Coefficients of variation for intra-observer variation were 7% and 14% for microCT and caliper measurements, respectively. Regression coefficients between observers were 0.97 for microCT and 0.91 for caliper measurements. Conclusion: MicroCT was more accurate than both caliper and F-FDG-PET for in vivo volumetric measurements of subcutaneous tumors in mice. F-FDG-PET was considered unsuitable for determination of tumor size. External caliper were inaccurate and encumbered with a significant and size dependent bias. MicroCT was also the most reproducible of the methods. Page 1 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Volume Determination Background Measurement of tumor size is important in preclinical Three weeks after implantation of tumor cells (tumor size ) volumes of 20 tumors were determined in animal studies when assessing responses to cancer treat- 20 – 250 mm ment. In longitudinal studies, sequential measurements vivo by external caliper, microCT and F-FDG-PET. Subse- of tumor volume with a non-invasive method are essen- quently, tumors were excised and reference tumor volume tial. Current standard technique for volume determina- was calculated from weight and density (1.05 g/mL). tion of subcutaneously xenografted tumors in vivo is by external caliper where tumor volume is calculated by use In order to assess the intra- and inter-observer variation of the modified ellipsoid formula 1/2(Length × on the microCT and caliper volume measurements two Width )[1,2]. However, measurements using caliper are additional experiments were carried out. To determine the often affected by errors due to e.g. variability in tumor intra-observer reproducibility, volume of two tumors was shape, skin thickness and subcutaneous fat layer thick- determined by acquiring 10 microCT scans and 10 caliper ness. Furthermore, observer subjectivity and differences in determinations of each tumor. In addition, volumes of a the compressibility of the tumor can easily lead to varia- group of 10 tumors were determined independently by tion in measurements. Clinically, computed tomography two different observers to assess inter-observer variation. (CT) and positron emission tomography (PET) are widely Each of the 10 mice had one microCT scan performed and used to monitor response to treatment [3]. Preclinical region of interests (ROIs) were subsequently drawn cover- imaging with microCT and microPET has in recent years ing the tumors independently by each of the two observ- become more widespread [4-8]. ers. The aim of the present study was therefore to evaluate if Measurement by Caliper microCT gives more accurate and reproducible measures In order to determine tumor volume by external caliper, of tumor volume in in vivo studies of subcutaneous the greatest longitudinal diameter (length) and the great- xenografted tumors compared with standard caliper est transverse diameter (width) were determined. Tumor measurements. Furthermore, we evaluated the accuracy of volume based on caliper measurements were calculated tumor volume determined from F-fluorodeoxyglucose by the modified ellipsoidal formula [1,2] ( F-FDG) PET. To do so, we compared the microCT, PET and caliper methods for tumor volume determination, Tumor volume = 1/2(length × width ) with ex vivo measurements as reference, and quantified inter- and intra-observer variation for the microCT and F-FDG microPET imaging and microCT imaging caliper methods. Mice were injected i.p or i.v with 8.7 ± 1.7 (mean ± SD) 18 18 MBq of F-FDG. F-FDG was produced at our own facil- ities (Rigshospitalet, Copenhagen, Denmark). One hour Methods Tumor Model after F-FDG injection mice were anaesthetized with 3% Six weeks old female NMRI (Naval Medical Research Insti- sevofluran (Abbott Scandinavia AB, Solna, Sweden) tute) nude mice were acquired from Taconic Europe (Lille mixed with 35% O in N and fixed on a bed. A 20 min 2 2 Skensved, Denmark) and allowed to acclimate one week PET scan was acquired using a MicroPET Focus 120 (Sie- in the animal facility before any intervention was initi- mens Medical Solutions, Malvern, PA, USA). After data ated. All experimental procedures were conducted with acquisition, PET data were arranged into sinograms and the guidelines set forth by the Danish Ministry of Justice. subsequently reconstructed with the Ordered Subset Estrogen pellets, 0.72 mg 17-β-Estradiol, 60-day release Expectation Maximization 2D (OSEM2D) reconstruction (Innovative Research of America, Sarasota, FL, USA), were algorithm. The pixel size was 0.866 × 0.866 × 0.796 mm implanted s.c. during anesthesia with 1:1 v/v mixture of and in the center field of view the resolution was 1.4 mm Hypnorm (Janssen Pharmaceutica, Beerse, Belgium) and full-width-at-half-maximum. Dormicum (Roche, Basel, Switzerland). One week after implantation of pellets, MCF-7 (human breast adenocar- Following the microPET scan, a microCT scan was 7 ® cinoma) tumor cells (10 cells in 100 μL medium mixed acquired with a MicroCAT II system (Siemens Medical with 100 μL Matrixgel™ Basement Membrane Matrix (BD solutions). A 7 minute and 10 seconds CT scan was per- Biosciences, San Jose, CA, USA)) were injected subcutane- formed with parameter settings: 360 rotation steps, tube ous into the left and right flank respectively. Cells were voltage 60 kV, tube current 500 μA, binning 4 and expo- cultured in Dulbecco's Modified Eagle Medium (DMEM) sure time 310 ms. The pixel size was 0.081 × 0.081 × 0.081 medium supplemented with 10% fetal calf serum and 1% mm. penicillin-streptomycin in 5% CO at 37°C. PET and microCT images were separately analyzed with the Inveon software (Siemens). ROIs were drawn manu- Page 2 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 ally by qualitative assessment covering the entire tumor had systematic bias when compared to reference volume. and tumor volume was generated by summation of voxels The average overestimate of volume using PET was 52 mm (35%) and overestimation of volume by using cali- within the tomographic planes. per was 95 mm (86%). Statistical analysis Analysis of tumor volumes obtained by the three different Intra- and inter-observer variation methods was performed by linear regression for each Intra-observer variation expressed as CV was on average 3 3 method against reference tumor volume. Agreement 7% (83.4 ± 6.9 mm , n = 10; 93.4 ± 6.1 mm , n = 10; between the PET, microCT and caliper methods against mean ± SD) for the microCT method and 14% (102.4 ± 3 3 reference tumor volume was further analyzed by means of 17.4 mm , n = 10; 167.3 ± 18.0 mm , n = 10; mean ± SD) Bland-Altman plots where the central line (mean) indi- for the caliper method. cates the bias and the outer lines (± 2SD) indicate the lim- its of agreements (LoA) [9]. The 95% confidence interval Inter-observer variation expressed as correlations of (CI) on bias was calculated, and bias was considered sig- tumor volume determined by the two observers are ) were 0.97 nificant if 0 was not included in the CI. shown in figure 4. Correlation coefficients (R for microCT and 0.91 for caliper measurements (n = 10). In order to assess the intra-observer variation, coefficient Bland-Altman plots of the difference between two observ- of variation (CV) for the 10 repeated microCT and caliper ers against mean volume are shown in figure 5. measurements were calculated. Tumor volumes measured by two experienced observers were evaluated my means of The mean difference between observers for caliper meas- linear regression, correlation coefficients and Bland-Alt- urements was -9.8 mm (95% CI on difference: -24.5-4.9 3 3 man plots in order to determine the inter-observer varia- mm ; LoA: -56.3-36.7 mm ) and for microCT measure- tion. ments it was 0.0 mm (95% CI on difference: -5.9-5.9 3 3 mm ; LoA: -18.6-18.6 mm ). Accordingly, no systematic bias was found between observers. Results Tumor volume determined by microCT, F-FDG-PET and external caliper Discussion Tumor volume measured by microCT, PET and caliper all Most cancer treatment studies assess drug effect by correlated (P < 0.001) with reference volume (figure 1). sequential measurements of tumor volume. Currently, the MicroCT versus reference volume (n = 20) had the best fit standard method for non-invasive volume measurements of line y = 1.01 ± 0.04x - 6.1 ± 6.3 (R = 0.97; p < 0.001). of subcutaneous tumors in mice is with external caliper. Since 2 tumors were unidentifiable on the PET scan only This is a somewhat subjective method often affected by 18 tumors were available for comparison of F-FDG-PET much error [1] and accordingly there is a need for more with reference volume. The tumors used in the study did accurate volume measurements. Non-invasive imaging not contain necrotic elements. The best line for F-FDG- modalities such as ultrasonography and MRI have been PET versus reference volume was y = 1.24 ± 0.18x + 15.4 ± investigated for their ability to follow tumors in mice dur- 30.4 (R = 0.75; p < 0.001). Caliper versus reference vol- ing longitudinal studies in vivo [10,11]. Ultrasonography ume (n = 20) had the best fit of line y = 1.27 ± 0.15x + 56.9 and MRI were both shown to be valuable tools for esti- = 0.80; p < 0.001). An example of ROIs drawn ± 24.2 (R mating volume of small tumor masses. Ultrasonography in the microCT and PET pictures is shown in figure 2. has the advantage of having a relative low cost of equip- ment and both MRI and ultrasonography have the advan- Bland-Altman plots of volume measured by microCT, PET tage that they do not impose any radiation dose that may and caliper versus reference tumor volume are shown in interfere with tumor growth. Further, it has been shown figure 3. that CT as part of a clinical PET/CT scanner can determine volume more precisely than traditional caliper measure- The mean difference between microCT and reference ments of large subcutaneous tumors in rats [12]. Preclini- tumor volume was -5.1 mm (95% CI on difference: - cal imaging of small animals with dedicated animal 3 3 10.6-0.3 mm ; LoA: -29.6-19.4 mm ). Accordingly, no sig- microCT and microPET scanners has in recent years nificant bias of microCT measurements compared with become available and could be even better alternatives for the reference was found. The mean difference between tumor volume determination. Accordingly, we evaluated PET and reference tumor volume was 52.1 mm (95% CI the capability of these modalities for volume determina- 3 3 on difference: 28.0–73.3 mm ; LoA: -50.4–154.7 mm ) tion of subcutaneous tumors in mice. and between caliper and reference tumor volume it was 3 3 95.1 mm (95% CI on difference: 71.6–118.6 mm ; LoA: We found that the microCT method was altogether more -10.0–200.3 mm ). Accordingly, both PET and caliper accurate than both PET and caliper methods for determi- Page 3 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 0 50 100 150 200 250 300 0 50 100 150 200 250 300 Reference tumor volume (mm ) Reference tumor volume (mm ) 0 50 100 150 200 250 300 Reference tumor volume (mm ) Figure 1 Linear regression for microCT, F-FDG-PET and caliper determined tumor volume against reference tumor volume Linear regression for microCT, F-FDG-PET and caliper determined tumor volume against reference tumor 2 2 volume. Best lines were: y = 1.01x - 6.1 (R = 0.97) for microCT versus reference volume, y = 1.24x + 15.4 (R = 0.75) for 18 2 F-FDG-PET versus reference volume and y = 1.27x + 56.9 (R = 0.80) for caliper versus reference volume. nation of tumor volume. For the microCT method there quently, volume changes measured with caliper in small was no systematic bias, whereas both the PET and caliper and large tumors are not comparable and effects of anti- method systematically overestimated tumor size. The bias cancer drugs can easily be missed as tumors will tend of the caliper measurements was smaller for small tumors towards being determined with a greater bias as they grow compared to greater tumors also relatively seen. Conse- Page 4 of 9 (page number not for citation purposes) 3 3 CT tumor volume (mm) Caliper tumor volume (mm) PET tumor volume (mm) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 A B Tumor Tumor Figure 2 Transverse section of a representative F-FDG-PET (A) and microCT (B) image of a mouse with a subcutaneous tumor Transverse section of a representative F-FDG-PET (A) and microCT (B) image of a mouse with a subcutane- ous tumor. Tumor is indicated by a white arrow and ROIs are drawn separately in the PET and microCT picture. larger. The bias in PET measurements was not dependent Tumor volume was determined with a rather high repro- on tumor size. ducibility between observers by both the microCT and cal- iper methods. Variation between observers for microCT Intra-observer variation on the microCT measurements measurements in this study was comparable to a study of was substantially lower compared to variation on the cal- much greater tumors in rats [12], however the current iper measurements. In consequence, microCT measure- study has showed that the low inter-observer variation of ments will allow for detection of smaller changes and the microCT method is also valid for small tumors. earlier recognition of efficacy in subcutaneous xenografts during experimental cancer treatment studies than stand- Volume determination by the caliper method was inaccu- ard external caliper. Further, it will allow for reduction in rate with a significant bias that increased with tumor size. the number of animals necessary to show a given effect in Very likely, this inaccuracy is partly due to the assumption cancer treatment studies e.g. when testing new anticancer that all tumors have shape like a modified ellipsoid, drugs. which may be less true for large tumors. With the microCT method, bias that arises from assumption of this specific Previous, analysis of intra-observer variation of caliper geometry is removed. In consequence, tumor volume is measurements have been carried out. CV was 12% for a accurately determined irrespective of tumor size and form. 3 3 small (320 mm ) and 6% for a larger (1450 mm ) tumor [1]. Using even smaller tumors (70–90 mm ) we found Tumor volume measured by microPET did not correlate CV to be 14% for the caliper method and accordingly well with true tumor volume. In order to determine tumor overlooking effects during longitudinal treatment studies size by F-FDG-PET, all parts of the tumor must take up can be marked with this method. In contrast, in the F-FDG. Visual inspection of PET images in this study present study, where small tumors were used for the intra- (figure 2) showed a heterogeneous F-FDG uptake in the observer variation, we fund a CV for microCT measure- tumors, which made it difficult accurately to identify ments as low as 7%. MicroCT hence allows detection of tumor boundary. The resolution of the microPET images small changes in much smaller tumors than the tradi- was not as high as of the microCT images which also con- tional caliper. tributed to the much lower accuracy of the PET volume data. Therefore, it was not unexpected that F-FDG-PET Page 5 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Mean + 2SD Mean -10 -20 -30 Mean - 2SD -40 0 50 100 150 200 250 300 350 Mean of CT and reference volume (mm ) Mean + 2SD 50 Mean -50 Mean - 2SD -100 0 50 100 150 200 250 300 350 400 Mean of PET and reference volume (mm ) 200 Mean + 2SD Mean Mean - 2SD -50 0 50 100 150 200 250 300 350 400 Mean of caliper and reference volume (mm ) B re Figure 3 la fer nd- en Altman plots compar ce volume ing three methods for measurement of tumor volume of subcutaneous mouse xenografts with the Bland-Altman plots comparing three methods for measurement of tumor volume of subcutaneous mouse xenografts with the reference volume. The central line (mean) indicates the bias and the outer lines (± 2SD) indicate the limits of agreement (LoA). Page 6 of 9 (page number not for citation purposes) 3 3 3 Difference in volume (mm) Difference in volume (mm) Difference in volume (mm) caliper-reference PET-reference CT-reference BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 0 50 100 150 200 250 300 350 0 50 100 150 200 250 Observer 1 - caliper tumor volume (mm ) Observer 1 - CT tumor volume (mm ) Figure 4 Correlation of tumor volume determined by two different observers measured by caliper and microCT respectively Correlation of tumor volume determined by two different observers measured by caliper and microCT respectively. Tumor volumes measures by the two observers were plotted and correlations were evaluated by means of lin- 2 2 2 ear fitting and correlation coefficients (R ). Best line was y = 0.98x - 7.4 (R = 0.91) for caliper and y = 1.01x - 0.65 (R = 0.97) for microCT measurements. was unsuitable for determination of tumor volume and ible than caliper measurements. Consequently, microCT consequently F-FDG-PET is rarely used for volume is a promising method that should be used when studies measurements. As the current study showed that the of small changes in experimental cancer treatment studies microCT method accurately and precisely identified of subcutaneous tumors in mice is needed. tumor volumes, identification of tumors based on the anatomically CT image and subsequently fusion of PET Abbreviations and CT images will allow much more precise determina- CI: Confidence interval; CT: Computed tomography; CV: 18 18 tion of tracer uptake. Accordingly, a combination of Coefficient of variation; F-FDG: F-fluorodeoxyglucose; LoA: Limits of agreement; PET: Positron emission tomog- microCT with microPET will allow a sensitive and accu- rate quantification of tumor burden in mice and be valu- raphy; ROI: Region of interest; SUV: Standard uptake able for the evaluation of novel cancer treatments. value. Conclusion Competing interests In summary, the present study demonstrated that The authors declare that they have no competing interests. microCT was more accurate than both external caliper measurements and F-FDG-microPET for in vivo volu- Authors' contributions metric measurements of subcutaneous tumors in MMJ: conception and design of the study, animal studies, mice. F-FDG-microPET was considered unsuitable for image and data analysis, draft of manuscript. JTJ and TB: determination of tumor size. External caliper were inaccu- animal studies and image analysis. AK: conception and rate and encumbered with a significant and size depend- design of the study, draft of manuscript. All authors read ent bias. External caliper are, despite this inaccuracy, and approved the final manuscript. currently the standard method for determination of tumor volume due to the low cost and high throughput of the simple method. In contrast, we found that microCT was accurate, without systematic bias and more reproduc- Page 7 of 9 (page number not for citation purposes) Observer 2 - caliper tumor volume (mm ) Observer 2 - CT tumor volume (mm ) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Caliper Mean + 2SD Mean -20 -40 Mean - 2SD -60 -80 0 50 100 150 200 250 300 350 400 Mean of volume determined by observer 1 and 2 (mm ) CT Mean + 2SD Mean -5 -10 -15 Mean - 2SD -20 -25 0 50 100 150 200 250 300 Mean of volume determined by observer 1 and 2 (mm ) Bland-Altman plots of th Figure 5 e difference between two observers against mean tumor volume Bland-Altman plots of the difference between two observers against mean tumor volume. The central line (mean) indicates the bias and the outer lines (± 2SD) indicate the limits of agreement (LoA). Page 8 of 9 (page number not for citation purposes) 3 3 Difference in volume (mm) Difference in volume (mm) observer 1-2 observer 1-2 BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Acknowledgements Financial support for the study from the Danish National Advanced Tech- nology Foundation, the AP Moeller Foundation, the Novo Nordic Founda- tion and the Lundbeck Foundation is gratefully acknowledged. References 1. Euhus DM, Hudd C, LaRegina MC, Johnson FE: Tumor measure- ment in the nude mouse. J Surg Oncol 1986, 31:229-234. 2. Tomayko MM, Reynolds CP: Determination of subcutaneous tumor size in athymic (nude) mice, Cancer Chemother. Pharmacol 1989, 24:148-154. 3. Weber WA, Wieder H: Monitoring chemotherapy and radio- therapy of solid tumors, Eur. J Nucl Med Mol Imaging 2006, 33(Suppl 1):27-37. 4. Dorow DS, Cullinane C, Conus N, Roselt P, Binns D, McCarthy TJ, McArthur GA, Hicks RJ: Multi-tracer small animal PET imaging of the tumour response to the novel pan-Erb-B inhibitor CI- 1033. Eur J Nucl Med Mol Imaging 2006, 33:441-452. 5. Leyton J, Alao JP, Da CM, Stavropoulou AV, Latigo JR, Perumal M, Pil- lai R, He Q, Atadja P, Lam EW, Workman P, Vigushin DM, Aboagye EO: In vivo biological activity of the histone deacetylase inhibitor LAQ824 is detectable with 3'-deoxy-3'-[18F]fluor- othymidine positron emission tomography. Cancer Res 2006, 66:7621-7629. 6. Molthoff CF, Klabbers BM, Berkhof J, Felten JT, van Gelder M, Wind- horst AD, Slotman BJ, Lammertsma AA: Monitoring response to radiotherapy in human squamous cell cancer bearing nude mice: comparison of 2'-deoxy-2'-[18F]fluoro-D-glucose (FDG) and 3'-[18F]fluoro-3'-deoxythymidine (FLT). Mol Imaging Biol 2007, 9:340-347. 7. Su H, Bodenstein C, Dumont RA, Seimbille Y, Dubinett S, Phelps ME, Herschman H, Czernin J, Weber W: Monitoring tumor glucose utilization by positron emission tomography for the predic- tion of treatment response to epidermal growth factor receptor kinase inhibitors. Clin Cancer Res 2006, 12:5659-5667. 8. Waldherr C, Mellinghoff IK, Tran C, Halpern BS, Rozengurt N, Safaei A, Weber WA, Stout D, Satyamurthy N, Barrio J, Phelps ME, Silver- man DH, Sawyers CL, Czernin J: Monitoring antiproliferative responses to kinase inhibitor therapy in mice with 3'-deoxy- 3'-18F-fluorothymidine PET. J Nucl Med 2005, 46:114-120. 9. Bland JM, Altman DG: Statistical methods for assessing agree- ment between two methods of clinical measurement. Lancet 1986, 1:307-310. 10. Cheung AM, Brown AS, Hastie LA, Cucevic V, Roy M, Lacefield JC, Fenster A, Foster FS: Three-dimensional ultrasound biomicro- scopy for xenograft growth analysis. Ultrasound Med Biol 2005, 31:865-870. 11. Mazurchuk R, Glaves D, Raghavan D: Magnetic resonance imag- ing of response to chemotherapy in orthotopic xenografts of human bladder cancer. Clin Cancer Res 1997, 3:1635-1641. 12. Ishimori T, Tatsumi M, Wahl RL: Tumor response assessment is more robust with sequential CT scanning than external cal- iper measurements. Acad Radiol 2005, 12:776-781. Pre-publication history The pre-publication history for this paper can be accessed Publish with Bio Med Central and every here: scientist can read your work free of charge http://www.biomedcentral.com/1471-2342/8/16/prepub "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 9 of 9 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png BMC Medical Imaging Springer Journals

Tumor volume in subcutaneous mouse xenografts measured by microCT is more accurate and reproducible than determined by 18F-FDG-microPET or external caliper

Download PDF
9 pages

Loading next page...
 
/lp/springer-journals/tumor-volume-in-subcutaneous-mouse-xenografts-measured-by-microct-is-J3I9evv0Ii

References (14)

Publisher
Springer Journals
Copyright
Copyright © 2008 by Jensen et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Imaging / Radiology
eISSN
1471-2342
DOI
10.1186/1471-2342-8-16
pmid
18925932
Publisher site
See Article on Publisher Site

Abstract

Background: In animal studies tumor size is used to assess responses to anticancer therapy. Current standard for volumetric measurement of xenografted tumors is by external caliper, a method often affected by error. The aim of the present study was to evaluate if microCT gives more accurate and reproducible measures of tumor size in mice compared with caliper measurements. Furthermore, we evaluated the accuracy of tumor volume determined from F- fluorodeoxyglucose ( F-FDG) PET. Methods: Subcutaneously implanted human breast adenocarcinoma cells in NMRI nude mice served as tumor model. Tumor volume (n = 20) was determined in vivo by external caliper, microCT and F-FDG-PET and subsequently reference volume was determined ex vivo. Intra- observer reproducibility of the microCT and caliper methods were determined by acquiring 10 repeated volume measurements. Volumes of a group of tumors (n = 10) were determined independently by two observers to assess inter-observer variation. Results: Tumor volume measured by microCT, PET and caliper all correlated with reference volume. No significant bias of microCT measurements compared with the reference was found, whereas both PET and caliper had systematic bias compared to reference volume. Coefficients of variation for intra-observer variation were 7% and 14% for microCT and caliper measurements, respectively. Regression coefficients between observers were 0.97 for microCT and 0.91 for caliper measurements. Conclusion: MicroCT was more accurate than both caliper and F-FDG-PET for in vivo volumetric measurements of subcutaneous tumors in mice. F-FDG-PET was considered unsuitable for determination of tumor size. External caliper were inaccurate and encumbered with a significant and size dependent bias. MicroCT was also the most reproducible of the methods. Page 1 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Volume Determination Background Measurement of tumor size is important in preclinical Three weeks after implantation of tumor cells (tumor size ) volumes of 20 tumors were determined in animal studies when assessing responses to cancer treat- 20 – 250 mm ment. In longitudinal studies, sequential measurements vivo by external caliper, microCT and F-FDG-PET. Subse- of tumor volume with a non-invasive method are essen- quently, tumors were excised and reference tumor volume tial. Current standard technique for volume determina- was calculated from weight and density (1.05 g/mL). tion of subcutaneously xenografted tumors in vivo is by external caliper where tumor volume is calculated by use In order to assess the intra- and inter-observer variation of the modified ellipsoid formula 1/2(Length × on the microCT and caliper volume measurements two Width )[1,2]. However, measurements using caliper are additional experiments were carried out. To determine the often affected by errors due to e.g. variability in tumor intra-observer reproducibility, volume of two tumors was shape, skin thickness and subcutaneous fat layer thick- determined by acquiring 10 microCT scans and 10 caliper ness. Furthermore, observer subjectivity and differences in determinations of each tumor. In addition, volumes of a the compressibility of the tumor can easily lead to varia- group of 10 tumors were determined independently by tion in measurements. Clinically, computed tomography two different observers to assess inter-observer variation. (CT) and positron emission tomography (PET) are widely Each of the 10 mice had one microCT scan performed and used to monitor response to treatment [3]. Preclinical region of interests (ROIs) were subsequently drawn cover- imaging with microCT and microPET has in recent years ing the tumors independently by each of the two observ- become more widespread [4-8]. ers. The aim of the present study was therefore to evaluate if Measurement by Caliper microCT gives more accurate and reproducible measures In order to determine tumor volume by external caliper, of tumor volume in in vivo studies of subcutaneous the greatest longitudinal diameter (length) and the great- xenografted tumors compared with standard caliper est transverse diameter (width) were determined. Tumor measurements. Furthermore, we evaluated the accuracy of volume based on caliper measurements were calculated tumor volume determined from F-fluorodeoxyglucose by the modified ellipsoidal formula [1,2] ( F-FDG) PET. To do so, we compared the microCT, PET and caliper methods for tumor volume determination, Tumor volume = 1/2(length × width ) with ex vivo measurements as reference, and quantified inter- and intra-observer variation for the microCT and F-FDG microPET imaging and microCT imaging caliper methods. Mice were injected i.p or i.v with 8.7 ± 1.7 (mean ± SD) 18 18 MBq of F-FDG. F-FDG was produced at our own facil- ities (Rigshospitalet, Copenhagen, Denmark). One hour Methods Tumor Model after F-FDG injection mice were anaesthetized with 3% Six weeks old female NMRI (Naval Medical Research Insti- sevofluran (Abbott Scandinavia AB, Solna, Sweden) tute) nude mice were acquired from Taconic Europe (Lille mixed with 35% O in N and fixed on a bed. A 20 min 2 2 Skensved, Denmark) and allowed to acclimate one week PET scan was acquired using a MicroPET Focus 120 (Sie- in the animal facility before any intervention was initi- mens Medical Solutions, Malvern, PA, USA). After data ated. All experimental procedures were conducted with acquisition, PET data were arranged into sinograms and the guidelines set forth by the Danish Ministry of Justice. subsequently reconstructed with the Ordered Subset Estrogen pellets, 0.72 mg 17-β-Estradiol, 60-day release Expectation Maximization 2D (OSEM2D) reconstruction (Innovative Research of America, Sarasota, FL, USA), were algorithm. The pixel size was 0.866 × 0.866 × 0.796 mm implanted s.c. during anesthesia with 1:1 v/v mixture of and in the center field of view the resolution was 1.4 mm Hypnorm (Janssen Pharmaceutica, Beerse, Belgium) and full-width-at-half-maximum. Dormicum (Roche, Basel, Switzerland). One week after implantation of pellets, MCF-7 (human breast adenocar- Following the microPET scan, a microCT scan was 7 ® cinoma) tumor cells (10 cells in 100 μL medium mixed acquired with a MicroCAT II system (Siemens Medical with 100 μL Matrixgel™ Basement Membrane Matrix (BD solutions). A 7 minute and 10 seconds CT scan was per- Biosciences, San Jose, CA, USA)) were injected subcutane- formed with parameter settings: 360 rotation steps, tube ous into the left and right flank respectively. Cells were voltage 60 kV, tube current 500 μA, binning 4 and expo- cultured in Dulbecco's Modified Eagle Medium (DMEM) sure time 310 ms. The pixel size was 0.081 × 0.081 × 0.081 medium supplemented with 10% fetal calf serum and 1% mm. penicillin-streptomycin in 5% CO at 37°C. PET and microCT images were separately analyzed with the Inveon software (Siemens). ROIs were drawn manu- Page 2 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 ally by qualitative assessment covering the entire tumor had systematic bias when compared to reference volume. and tumor volume was generated by summation of voxels The average overestimate of volume using PET was 52 mm (35%) and overestimation of volume by using cali- within the tomographic planes. per was 95 mm (86%). Statistical analysis Analysis of tumor volumes obtained by the three different Intra- and inter-observer variation methods was performed by linear regression for each Intra-observer variation expressed as CV was on average 3 3 method against reference tumor volume. Agreement 7% (83.4 ± 6.9 mm , n = 10; 93.4 ± 6.1 mm , n = 10; between the PET, microCT and caliper methods against mean ± SD) for the microCT method and 14% (102.4 ± 3 3 reference tumor volume was further analyzed by means of 17.4 mm , n = 10; 167.3 ± 18.0 mm , n = 10; mean ± SD) Bland-Altman plots where the central line (mean) indi- for the caliper method. cates the bias and the outer lines (± 2SD) indicate the lim- its of agreements (LoA) [9]. The 95% confidence interval Inter-observer variation expressed as correlations of (CI) on bias was calculated, and bias was considered sig- tumor volume determined by the two observers are ) were 0.97 nificant if 0 was not included in the CI. shown in figure 4. Correlation coefficients (R for microCT and 0.91 for caliper measurements (n = 10). In order to assess the intra-observer variation, coefficient Bland-Altman plots of the difference between two observ- of variation (CV) for the 10 repeated microCT and caliper ers against mean volume are shown in figure 5. measurements were calculated. Tumor volumes measured by two experienced observers were evaluated my means of The mean difference between observers for caliper meas- linear regression, correlation coefficients and Bland-Alt- urements was -9.8 mm (95% CI on difference: -24.5-4.9 3 3 man plots in order to determine the inter-observer varia- mm ; LoA: -56.3-36.7 mm ) and for microCT measure- tion. ments it was 0.0 mm (95% CI on difference: -5.9-5.9 3 3 mm ; LoA: -18.6-18.6 mm ). Accordingly, no systematic bias was found between observers. Results Tumor volume determined by microCT, F-FDG-PET and external caliper Discussion Tumor volume measured by microCT, PET and caliper all Most cancer treatment studies assess drug effect by correlated (P < 0.001) with reference volume (figure 1). sequential measurements of tumor volume. Currently, the MicroCT versus reference volume (n = 20) had the best fit standard method for non-invasive volume measurements of line y = 1.01 ± 0.04x - 6.1 ± 6.3 (R = 0.97; p < 0.001). of subcutaneous tumors in mice is with external caliper. Since 2 tumors were unidentifiable on the PET scan only This is a somewhat subjective method often affected by 18 tumors were available for comparison of F-FDG-PET much error [1] and accordingly there is a need for more with reference volume. The tumors used in the study did accurate volume measurements. Non-invasive imaging not contain necrotic elements. The best line for F-FDG- modalities such as ultrasonography and MRI have been PET versus reference volume was y = 1.24 ± 0.18x + 15.4 ± investigated for their ability to follow tumors in mice dur- 30.4 (R = 0.75; p < 0.001). Caliper versus reference vol- ing longitudinal studies in vivo [10,11]. Ultrasonography ume (n = 20) had the best fit of line y = 1.27 ± 0.15x + 56.9 and MRI were both shown to be valuable tools for esti- = 0.80; p < 0.001). An example of ROIs drawn ± 24.2 (R mating volume of small tumor masses. Ultrasonography in the microCT and PET pictures is shown in figure 2. has the advantage of having a relative low cost of equip- ment and both MRI and ultrasonography have the advan- Bland-Altman plots of volume measured by microCT, PET tage that they do not impose any radiation dose that may and caliper versus reference tumor volume are shown in interfere with tumor growth. Further, it has been shown figure 3. that CT as part of a clinical PET/CT scanner can determine volume more precisely than traditional caliper measure- The mean difference between microCT and reference ments of large subcutaneous tumors in rats [12]. Preclini- tumor volume was -5.1 mm (95% CI on difference: - cal imaging of small animals with dedicated animal 3 3 10.6-0.3 mm ; LoA: -29.6-19.4 mm ). Accordingly, no sig- microCT and microPET scanners has in recent years nificant bias of microCT measurements compared with become available and could be even better alternatives for the reference was found. The mean difference between tumor volume determination. Accordingly, we evaluated PET and reference tumor volume was 52.1 mm (95% CI the capability of these modalities for volume determina- 3 3 on difference: 28.0–73.3 mm ; LoA: -50.4–154.7 mm ) tion of subcutaneous tumors in mice. and between caliper and reference tumor volume it was 3 3 95.1 mm (95% CI on difference: 71.6–118.6 mm ; LoA: We found that the microCT method was altogether more -10.0–200.3 mm ). Accordingly, both PET and caliper accurate than both PET and caliper methods for determi- Page 3 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 0 50 100 150 200 250 300 0 50 100 150 200 250 300 Reference tumor volume (mm ) Reference tumor volume (mm ) 0 50 100 150 200 250 300 Reference tumor volume (mm ) Figure 1 Linear regression for microCT, F-FDG-PET and caliper determined tumor volume against reference tumor volume Linear regression for microCT, F-FDG-PET and caliper determined tumor volume against reference tumor 2 2 volume. Best lines were: y = 1.01x - 6.1 (R = 0.97) for microCT versus reference volume, y = 1.24x + 15.4 (R = 0.75) for 18 2 F-FDG-PET versus reference volume and y = 1.27x + 56.9 (R = 0.80) for caliper versus reference volume. nation of tumor volume. For the microCT method there quently, volume changes measured with caliper in small was no systematic bias, whereas both the PET and caliper and large tumors are not comparable and effects of anti- method systematically overestimated tumor size. The bias cancer drugs can easily be missed as tumors will tend of the caliper measurements was smaller for small tumors towards being determined with a greater bias as they grow compared to greater tumors also relatively seen. Conse- Page 4 of 9 (page number not for citation purposes) 3 3 CT tumor volume (mm) Caliper tumor volume (mm) PET tumor volume (mm) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 A B Tumor Tumor Figure 2 Transverse section of a representative F-FDG-PET (A) and microCT (B) image of a mouse with a subcutaneous tumor Transverse section of a representative F-FDG-PET (A) and microCT (B) image of a mouse with a subcutane- ous tumor. Tumor is indicated by a white arrow and ROIs are drawn separately in the PET and microCT picture. larger. The bias in PET measurements was not dependent Tumor volume was determined with a rather high repro- on tumor size. ducibility between observers by both the microCT and cal- iper methods. Variation between observers for microCT Intra-observer variation on the microCT measurements measurements in this study was comparable to a study of was substantially lower compared to variation on the cal- much greater tumors in rats [12], however the current iper measurements. In consequence, microCT measure- study has showed that the low inter-observer variation of ments will allow for detection of smaller changes and the microCT method is also valid for small tumors. earlier recognition of efficacy in subcutaneous xenografts during experimental cancer treatment studies than stand- Volume determination by the caliper method was inaccu- ard external caliper. Further, it will allow for reduction in rate with a significant bias that increased with tumor size. the number of animals necessary to show a given effect in Very likely, this inaccuracy is partly due to the assumption cancer treatment studies e.g. when testing new anticancer that all tumors have shape like a modified ellipsoid, drugs. which may be less true for large tumors. With the microCT method, bias that arises from assumption of this specific Previous, analysis of intra-observer variation of caliper geometry is removed. In consequence, tumor volume is measurements have been carried out. CV was 12% for a accurately determined irrespective of tumor size and form. 3 3 small (320 mm ) and 6% for a larger (1450 mm ) tumor [1]. Using even smaller tumors (70–90 mm ) we found Tumor volume measured by microPET did not correlate CV to be 14% for the caliper method and accordingly well with true tumor volume. In order to determine tumor overlooking effects during longitudinal treatment studies size by F-FDG-PET, all parts of the tumor must take up can be marked with this method. In contrast, in the F-FDG. Visual inspection of PET images in this study present study, where small tumors were used for the intra- (figure 2) showed a heterogeneous F-FDG uptake in the observer variation, we fund a CV for microCT measure- tumors, which made it difficult accurately to identify ments as low as 7%. MicroCT hence allows detection of tumor boundary. The resolution of the microPET images small changes in much smaller tumors than the tradi- was not as high as of the microCT images which also con- tional caliper. tributed to the much lower accuracy of the PET volume data. Therefore, it was not unexpected that F-FDG-PET Page 5 of 9 (page number not for citation purposes) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Mean + 2SD Mean -10 -20 -30 Mean - 2SD -40 0 50 100 150 200 250 300 350 Mean of CT and reference volume (mm ) Mean + 2SD 50 Mean -50 Mean - 2SD -100 0 50 100 150 200 250 300 350 400 Mean of PET and reference volume (mm ) 200 Mean + 2SD Mean Mean - 2SD -50 0 50 100 150 200 250 300 350 400 Mean of caliper and reference volume (mm ) B re Figure 3 la fer nd- en Altman plots compar ce volume ing three methods for measurement of tumor volume of subcutaneous mouse xenografts with the Bland-Altman plots comparing three methods for measurement of tumor volume of subcutaneous mouse xenografts with the reference volume. The central line (mean) indicates the bias and the outer lines (± 2SD) indicate the limits of agreement (LoA). Page 6 of 9 (page number not for citation purposes) 3 3 3 Difference in volume (mm) Difference in volume (mm) Difference in volume (mm) caliper-reference PET-reference CT-reference BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 0 50 100 150 200 250 300 350 0 50 100 150 200 250 Observer 1 - caliper tumor volume (mm ) Observer 1 - CT tumor volume (mm ) Figure 4 Correlation of tumor volume determined by two different observers measured by caliper and microCT respectively Correlation of tumor volume determined by two different observers measured by caliper and microCT respectively. Tumor volumes measures by the two observers were plotted and correlations were evaluated by means of lin- 2 2 2 ear fitting and correlation coefficients (R ). Best line was y = 0.98x - 7.4 (R = 0.91) for caliper and y = 1.01x - 0.65 (R = 0.97) for microCT measurements. was unsuitable for determination of tumor volume and ible than caliper measurements. Consequently, microCT consequently F-FDG-PET is rarely used for volume is a promising method that should be used when studies measurements. As the current study showed that the of small changes in experimental cancer treatment studies microCT method accurately and precisely identified of subcutaneous tumors in mice is needed. tumor volumes, identification of tumors based on the anatomically CT image and subsequently fusion of PET Abbreviations and CT images will allow much more precise determina- CI: Confidence interval; CT: Computed tomography; CV: 18 18 tion of tracer uptake. Accordingly, a combination of Coefficient of variation; F-FDG: F-fluorodeoxyglucose; LoA: Limits of agreement; PET: Positron emission tomog- microCT with microPET will allow a sensitive and accu- rate quantification of tumor burden in mice and be valu- raphy; ROI: Region of interest; SUV: Standard uptake able for the evaluation of novel cancer treatments. value. Conclusion Competing interests In summary, the present study demonstrated that The authors declare that they have no competing interests. microCT was more accurate than both external caliper measurements and F-FDG-microPET for in vivo volu- Authors' contributions metric measurements of subcutaneous tumors in MMJ: conception and design of the study, animal studies, mice. F-FDG-microPET was considered unsuitable for image and data analysis, draft of manuscript. JTJ and TB: determination of tumor size. External caliper were inaccu- animal studies and image analysis. AK: conception and rate and encumbered with a significant and size depend- design of the study, draft of manuscript. All authors read ent bias. External caliper are, despite this inaccuracy, and approved the final manuscript. currently the standard method for determination of tumor volume due to the low cost and high throughput of the simple method. In contrast, we found that microCT was accurate, without systematic bias and more reproduc- Page 7 of 9 (page number not for citation purposes) Observer 2 - caliper tumor volume (mm ) Observer 2 - CT tumor volume (mm ) BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Caliper Mean + 2SD Mean -20 -40 Mean - 2SD -60 -80 0 50 100 150 200 250 300 350 400 Mean of volume determined by observer 1 and 2 (mm ) CT Mean + 2SD Mean -5 -10 -15 Mean - 2SD -20 -25 0 50 100 150 200 250 300 Mean of volume determined by observer 1 and 2 (mm ) Bland-Altman plots of th Figure 5 e difference between two observers against mean tumor volume Bland-Altman plots of the difference between two observers against mean tumor volume. The central line (mean) indicates the bias and the outer lines (± 2SD) indicate the limits of agreement (LoA). Page 8 of 9 (page number not for citation purposes) 3 3 Difference in volume (mm) Difference in volume (mm) observer 1-2 observer 1-2 BMC Medical Imaging 2008, 8:16 http://www.biomedcentral.com/1471-2342/8/16 Acknowledgements Financial support for the study from the Danish National Advanced Tech- nology Foundation, the AP Moeller Foundation, the Novo Nordic Founda- tion and the Lundbeck Foundation is gratefully acknowledged. References 1. Euhus DM, Hudd C, LaRegina MC, Johnson FE: Tumor measure- ment in the nude mouse. J Surg Oncol 1986, 31:229-234. 2. Tomayko MM, Reynolds CP: Determination of subcutaneous tumor size in athymic (nude) mice, Cancer Chemother. Pharmacol 1989, 24:148-154. 3. Weber WA, Wieder H: Monitoring chemotherapy and radio- therapy of solid tumors, Eur. J Nucl Med Mol Imaging 2006, 33(Suppl 1):27-37. 4. Dorow DS, Cullinane C, Conus N, Roselt P, Binns D, McCarthy TJ, McArthur GA, Hicks RJ: Multi-tracer small animal PET imaging of the tumour response to the novel pan-Erb-B inhibitor CI- 1033. Eur J Nucl Med Mol Imaging 2006, 33:441-452. 5. Leyton J, Alao JP, Da CM, Stavropoulou AV, Latigo JR, Perumal M, Pil- lai R, He Q, Atadja P, Lam EW, Workman P, Vigushin DM, Aboagye EO: In vivo biological activity of the histone deacetylase inhibitor LAQ824 is detectable with 3'-deoxy-3'-[18F]fluor- othymidine positron emission tomography. Cancer Res 2006, 66:7621-7629. 6. Molthoff CF, Klabbers BM, Berkhof J, Felten JT, van Gelder M, Wind- horst AD, Slotman BJ, Lammertsma AA: Monitoring response to radiotherapy in human squamous cell cancer bearing nude mice: comparison of 2'-deoxy-2'-[18F]fluoro-D-glucose (FDG) and 3'-[18F]fluoro-3'-deoxythymidine (FLT). Mol Imaging Biol 2007, 9:340-347. 7. Su H, Bodenstein C, Dumont RA, Seimbille Y, Dubinett S, Phelps ME, Herschman H, Czernin J, Weber W: Monitoring tumor glucose utilization by positron emission tomography for the predic- tion of treatment response to epidermal growth factor receptor kinase inhibitors. Clin Cancer Res 2006, 12:5659-5667. 8. Waldherr C, Mellinghoff IK, Tran C, Halpern BS, Rozengurt N, Safaei A, Weber WA, Stout D, Satyamurthy N, Barrio J, Phelps ME, Silver- man DH, Sawyers CL, Czernin J: Monitoring antiproliferative responses to kinase inhibitor therapy in mice with 3'-deoxy- 3'-18F-fluorothymidine PET. J Nucl Med 2005, 46:114-120. 9. Bland JM, Altman DG: Statistical methods for assessing agree- ment between two methods of clinical measurement. Lancet 1986, 1:307-310. 10. Cheung AM, Brown AS, Hastie LA, Cucevic V, Roy M, Lacefield JC, Fenster A, Foster FS: Three-dimensional ultrasound biomicro- scopy for xenograft growth analysis. Ultrasound Med Biol 2005, 31:865-870. 11. Mazurchuk R, Glaves D, Raghavan D: Magnetic resonance imag- ing of response to chemotherapy in orthotopic xenografts of human bladder cancer. Clin Cancer Res 1997, 3:1635-1641. 12. Ishimori T, Tatsumi M, Wahl RL: Tumor response assessment is more robust with sequential CT scanning than external cal- iper measurements. Acad Radiol 2005, 12:776-781. Pre-publication history The pre-publication history for this paper can be accessed Publish with Bio Med Central and every here: scientist can read your work free of charge http://www.biomedcentral.com/1471-2342/8/16/prepub "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 9 of 9 (page number not for citation purposes)

Journal

BMC Medical Imaging – Springer Journals

Published: Oct 16, 2008

There are no references for this article.