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

Learn More →

CanAssist Breast Impacting Clinical Treatment Decisions in Early-Stage HR+ Breast Cancer Patients: Indian Scenario

CanAssist Breast Impacting Clinical Treatment Decisions in Early-Stage HR+ Breast Cancer... CanAssist Breast (CAB) has thus far been validated on a retrospective cohort of 1123 patients who are mostly Indians. Distant metastasis–free survival (DMFS) of more than 95% was observed with significant separation (P < 0.0001) between low-risk and high-risk groups. In this study, we demonstrate the usefulness of CAB in guiding physicians to assess risk of cancer recurrence and to make informed treatment decisions for patients. Of more than 500 patients who have undergone CAB test, detailed analysis of 455 patients who were treated based on CAB-based risk predictions by more than 140 doctors across India is presented here. Majority of patients tested had node negative, T2, and grade 2 disease. Age and luminal subtypes did not affect the performance of CAB. On comparison with Adjuvant! Online (AOL), CAB categorized twice the number of patients into low risk indicating potential of overtreatment by AOL-based risk categorization. We assessed the impact of CAB testing on treatment decisions for 254 patients and observed that 92% low-risk patients were not given chemotherapy. Overall, we observed that 88% patients were either given or not given chemotherapy based on whether they were stratified as high risk or low risk for distant recurrence respectively. Based on these results, we conclude that CAB has been accepted by physicians to make treatment planning and provides a cost-effective alternative to other similar multigene prognostic tests currently available. . . . . Keywords CanAssist-Breast Decision impact Indian patients Prognostic Early-stage breast cancer Background is a composite measure of education, income, and life expec- tancy, MIR for a medium HDI country like India in 2016 was Among all cancers, breast cancer occupies the top position in found to be 0.5 compared with 0.19 for a very high HDI Indian women [1]. Based on projections, breast cancer patient country like the USA [4]. Besides economic reasons, Indian numbers could reach 1.9 million by 2020 [2]. Mortality to patients also present with more aggressive form of the disease, incidence rate (MIR) is a useful indicator of 5-year survival with an earlier age at onset and higher tumor stage and node across various cancers [3]. In a study stratifying MIR for positivity. Even with the presence of high risk factors, not breast cancer with Human Development Index (HDI) which every patient might benefit from chemotherapy and the chal- lenge is in identifying these patients accurately. In the past 60 years, since the first cancer patient enrolled * Manjiri M. Bakre in a randomized clinical trial conducted by the National manjiri@oncostemdiagnostics.com Surgical Adjuvant Breast and Bowel Project (NSABP) [5], there has been constant improvement in the effective OncoStem Diagnostics Private Limited, 4, Raja Ram Mohan Roy Road, Aanand Towers, 2nd Floor, Bangalore, Karnataka 560027, treatment and surgical management of breast cancer patients. India Starting with a combination of cyclophosphamide, Manipal Hospital, Bangalore, India methotrexate, and 5-flurouracil (CMF) in the 1950s, several 3 new drugs (anthracyclines and taxanes) and drug regimens Sri Shankara Cancer Hospital and Research Center, Bangalore, India have since evolved [6], in addition to hormone therapy, HCG, Bangalore, India for hormone receptor–positive breast cancer. While the Rajiv Gandhi Cancer Institute and Research Center, New Delhi, India efficacy improved in some cases, these drugs also produced Mangalore Institute of Oncology, Manipal, India unwanted side effects like cardiotoxicity and decrease in Father Muller Medical College, Mangalore, India blood cell counts. S22 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Optimal treatment helps in minimizing mortality and The Indian Council of Medical Research (ICMR) does not morbidity associated with the disease. Treating every include tests like Oncotype DX and MammaPrint in their rec- patient with the most aggressive form of treatment ommendations for treatment decision-making due to the lack may not always be productive, even if the patient does of data on Indian patients (https://www.icmr.nic.in/sites/ not have existing comorbidities. The Cochrane review default/files/guidelines/Breast_Cancer.pdf). The critical point compiled data from 14 randomized clinical trials involv- of development and validation on Indian patients makes ing 5600 women. There was quality data to show that CanAssist-Breast a good choice for Indian patients. high doses of chemotherapy did not improve survival in CanAssist Breast test has been used for treatment planning early-stage breast cancer patients [7]. The harmful side for patients since 2016. The aims of the present study are to effects of chemotherapy are well known and balancing briefly summarize the retrospective validation data about the harm to benefit from chemotherapy would improve CAB and analyze the clinical scenarios for recommending this the quality of life of patients. With the advent of test and correlation of risk stratification by CAB to treatment multigene prognostic tests, studies have shown 70% decisions by referring physicians. early-stage hormone receptor–positive, node-negative breast cancer patients benefit from chemotherapy [8]. Some of the commercially available prognostic tests in- Methods clude Oncotype DX (ODX) [9], MammaPrint [10], Prosigna [11], EndoPredict [12], Breast Cancer Index (BCI) [13], and Patient Selection CanAssist Breast (CAB) [14]. All these tests query different gene/protein markers and use different testing methodologies For the retrospective study, we obtained post-surgical tumor (Table 1) but have been shown to predict risk of recurrence samples in the form of formalin-fixed paraffin-embedded with more than 95% accuracy. (FFPE) blocks from 1123 early-stage breast cancer patients. Of all these tests, CAB is the only test that has been Patient consent, patient information, and treatment follow-up developed and validated on a mixed cohort of Asian details such as age, year of diagnosis, type of surgery, tumor (Indian) and Caucasian patients in a 3:1 proportion size and grade, hormone receptor status, node status, treatment [15] and extensively validated for analytical perfor- regimen, date of recurrence or last visit, or death were obtain- mance [16]. CAB is an immunohistochemistry (IHC)- ed from the treating hospitals. All patients had hormone based test that assesses risk of cancer recurrence at a receptor–positive disease. TNBC patients were excluded from distant site within 5 years from diagnosis. It quantifies the study. The patients were staged based on the AJCC 7th protein expression levels of a combination of 5 unique edition staging system. Patients with tumors with stage I non-proliferative biomarkers (CD44, Pan-Cadherin, N- (T1N0) and stage II (T1N1, T2N0, T2N1, T3N0) were con- Cadherin, ABCC4, and ABCC11). CAB markers are sidered as early stage. Patients with a minimum of 5-year involved in diverse cancer signaling pathways that reg- follow-up were included and this requirement was waived ulate cancer metastasis and drug resistance. The IHC off only in patients with a recurrence at a distant site within data from the biomarkers are combined with three clin- the 5-year period. ical parameters, tumor size (T), node status (N), and For the decision impact study, patients who were recom- tumor grade, to generate a low- or a high-risk score mended testing by CanAssist Breast by their treating physi- for every patient using a machine learning–based statis- cian between 2016 and May 2019 were included in this study. tical model. Patient consent was taken as part of test requisition form. Table 1 Summary of commercially available prognostic tests Oncotype DX MammaPrint Prosigna EndoPredict Breast CanAssist Breast Cancer Index No. of genes used 21 70 50 8 7 5 biomarkers Proliferation genes included Yes Yes Yes Yes No No Method qPCR DNA microarray NanoString qPCR qPCR IHC Clinical parameters No No No Tumor size, No Tumor size, node status node status, and tumor grade Presence of intermediate zone Yes No Yes No No No Prediction of chemotherapy benefit Yes No No No No Yes Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S23 Clinical information like node, tumor size, and tumor grade CAP-accredited central reference OncoStem lab at Bangalore, was obtained from pathology reports obtained from the pa- India. The IHC grading information was incorporated along tients along with the test requisition form. ER/PR status was with the 3 clinical parameters (tumor size, tumor grade, and included as a mandatory requirement for acceptance for nodal status) as obtained from the treating hospitals to calcu- CanAssist Breast testing and the report was provided by the late CAB risk scores on a scale of 0–100. A cutoff of 15.5 is patient or hospital from where the cases were referred by. used to classify patients into low risk (score ≤ 15.5) or high Ki67, wherever available along with the ER/PR reports, was risk (score > 15.5) for distant recurrence [14]. used for luminal classification in this study. Referring physi- cian name and patient age were obtained from the test requi- sition form. Luminal subtyping was performed as per St. Adjuvant! Online–Based Risk Categorization Gallen’s recommendations [17]. The modified Adjuvant! Online (ver8) criteria as described in Study Design the MINDACT trial [18] was used for assigning risk catego- ries based on tumor grade, node status, and tumor size A total of 455 patients who were recommended CAB for (Table 2). A total of 430 cases for which the exact T size and number of metastatic nodes data were available were used treatment planning were included for various subgroup anal- yses in this study. To assess region-wise distribution of pre- in this concordance analysis. scribing physicians from India, five different zones were con- sidered. These included North (Delhi, Uttar Pradesh, Haryana, and Rajasthan), Central (Madhya Pradesh), West Results (Maharashtra, Gujarat, and Rajasthan), East (West Begal), and South (Karnataka, Kerala, Andhra Pradesh, and Tamil Performance of CAB on Retrospective Cohort Nadu). Prescriptions were also received from neighboring countries of the Indian subcontinent comprising of Sri Retrospective patients (n = 1123) were dichotomized into Lanka, Pakistan, and Bangladesh which were considered as “outside India” for analysis. Information on usefulness of low- and high-risk groups by CAB. There was 11% difference in the DMFS between low- and high-risk groups demonstrat- CAB-based risk stratification for treatment planning was ob- tained from prescribing physicians via email and personal ing statistically significant (P < 0.0001) separation between the two groups (Fig. 1a). Since this cohort had a mix of pa- visits. Data was collected and analyzed to assess if patients tients with and without chemotherapy treatment, we further stratified as either low risk or high risk for recurrence by CAB performed survival analysis on patients who were treated with were given chemotherapy or not for patients who were pre- endocrine therapy alone (n = 298). This was to exclude any scribed CAB. This follow-up information was obtained for 254 patients. confounding effect of chemotherapy in the mixed cohort. The separation between the low- and high-risk groups was signif- Statistical Analysis icant (P = 0.0002) for this subgroup as well (Fig. 1b)with a clear difference in DMFS of 14% between the two risk groups. Kaplan-Meier survival curve analysis was performed using GraphPad version 8. Distant metastasis–free survival Table 2 Criteria used for clinical risk classification (DMFS) was calculated for CAB high-risk versus low-risk patients. P values were computed using log-rank two-sided Grade Nodal status Tumor size Clinical risk test at 0.05 significance. DMFS is the time interval between Well differentiated N0 ≤3cm C-Low the date of diagnosis of cancer and the last date of follow-up in (grade 1) 3.1–5cm C-High case of no event/recurrence with a minimum period of 5 years. N1 (1–3 ≤2cm C-Low MedCalc (https://www.medcalc.org/calc/comparison_of_ positive nodes) 2.1–5cm C-High proportions.php) was used to calculate statistical significance Moderately differentiated N0 ≤2cm C-Low of proportions. P values less than or equal to 0.05 were (grade 2) 2.1–5cm C-High considered to be statistically significant. N1 (1–3 Any size C-High positive nodes) CanAssist Breast Testing Poorly differentiated N0 ≤1cm C-Low (grade 3) 1.1–5cm C-High Testing was performed on FFPE blocks. The CAB was per- N1 (1–3 Any size C-High formed as described previously [14, 15]. Briefly, immunohis- positive nodes) tochemistry was performed for the 5 CAB biomarkers at the S24 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Fig. 1 Performance of CAB. a Survival (KM) curve using CAB- based risk categorization on a retrospective mixed cohort of chemotherapy-treated and chemotherapy-naïve (endocrine therapy alone treated) patients. b Survival analysis using CAB- based categorization with chemotherapy-naïve patient cohort The DMFS for both the mixed and chemotherapy-naïve sub- clinical characteristics were predominantly similar to that of groups was > 95% in the low-risk group. the retrospective cohort used for test validation. In the node-positive group of the prospective cohort, 94% of patients presented with N1 disease. Sixty-five percent of the Evaluation of Various Risk Factors in the Study Cohort patients with N1 disease were single-node positive, 16% were in the Prospective Cohort double-node positive, and 13% were triple-node positive. Based on the tumor sizes, 92% of T1 patients had a tumor Upon successful completion of clinical, analytical validation of size of ≥ 1 cm; 85% and 27% of T2 patients had tumor sizes CanAssist-Breast and upon getting the appropriate regulatory between 2–3and 3–4 cm respectively; 1.8% of patients had accreditations, CAB was launched in the market from the mid- tumor size > 5 cm. Percentage of grades 1 and 3 were 19.8 and dle of 2016. Until date, over 500 patients have availed CAB in 14.9% respectively. the Indian subcontinent. Analysis has been shown in the sub- 55.4% of patients were in the age group of 41–60 years and sequent sections on the prospective cohort of 455 patients. 5% patients had an age at onset of ≤ 40 years. 68.7% of pa- As shown in the retrospective clinical validation cohort data tients were of the luminal B subtype (Table 3). [15], we had observed that most patients had T2 tumors with grade 2 disease and node-negative disease [15]. We wanted to evaluate if this was the case with prospective patient cohort Clinical Factors Influencing CAB Risk Categorization who were prescribed CAB. We therefore performed subgroup analysis based on node status, tumor size, tumor grade, and We further evaluated the effect of the clinical factors on CAB- based risk categorization. Luminal subtype and age had no age. There was a high proportion of patients with N0 (82.4%), T2 (58.4%), and grade 2 (65.3%) disease (Table 3). Thus, significant effect on proportions of low-risk stratified patients by CAB, while the influence of tumor size, node, and tumor grade was significant (Table 4). We also evaluated the perfor- Table 3 Cohorts’ description (n = 455). For T size, n =448; luminal mance of the test across patients from two geographically subtyping, 265 patients for whom Ki67 status was known were distinct zones (North and South) of India and found that the considered proportions of low- and high-risk patients was not Parameter Number of patients (%) T1 178 (39.6) Table 4 Significance of low-risk proportions across various disease T2 262 (58.4) parameters T3 9 (2.0) Parameter % low risk P value N0 375 (82.4) N+ 80 (17.6) Luminal A 70 P =0.5 Luminal B 69 G1 90 (19.8) T1 83 P <0.0001 G2 297 (65.3) T2+T3 61 G3 68 (14.9) N0 74 P =0.0004 Luminal A 83 (31.3) N+ 54 Luminal B 182 (68.7) G1 89 P <0.0001 Age < 40 years 23 (5.0) G2+G3 65 Age 41–60 years 252 (55.4) <40 years 74 P =0.7 Age > 61 years 180 (39.6) >40 years 70 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S25 significantly different across these two patient subgroups (P = 2b). CAB categorized 34.4% N0 and 47.9% of N+ 0.8), and to the total pooled subgroup (P = 0.7 for South vs AOL (clinical) high-risk patients into low risk (Fig. 2c). total; P = 0.6 for North vs total). Comparison of CAB with Adjuvant! Online on Risk Physician Considerations for Testing Categorization The tests were prescribed by 147 physicians from all across In the scenario where patients cannot afford expensive India (North, South, Central, East, and West) and from neigh- prognostic tests, physicians tend to decide on therapy boring countries (Fig. 3a). While 47% of physicians pre- options based on various clinical parameters and freely scribed the test for more than one patient with 11% of physi- cians prescribing > 6 patients for CAB testing (Fig. 3b). Next, available online predictive tools like Predict [19], NPI [20], and Adjuvant! Online (AOL) [21]. While NPI and we wanted to assess the reasons for prescribing CAB by phy- sicians to their patients. From the survey conducted by us, the Predict provide estimate of overall survival, Adjuvant! Online provides risk of recurrence and potential re- major reasons were (a) presentation with high-risk clinical features like large tumors or node positivity, (b) to assess if sponse to chemotherapy. Since both CAB and Adjuvant! Online aid in making treatment decisions, patients with luminal A disease would benefit from chemo- therapy, (c) young age of patients who want to avoid side we compared the performance of CAB with AOL that predicts risk based on clinical parameters alone. For effects of chemotherapy, (d) old aged patients who might be those patients for whom we had information of the ex- spared chemotherapy if possible, and (e) comorbidities in the act size of the tumor (n = 430), we compared the risk patient that might require evaluating the benefit of chemother- categorization by AOL vs CAB. AOL (clinical) catego- apy over its effect that would have a bearing on the quality of life. Across all these reasons, the main rationale for prescrip- rized 62.4% patients into high risk as compared with 29% by CAB (Fig. 2a). CAB categorized 36.6% of tion for all the physicians was to avoid chemotherapy, if possible. AOL (clinical) high-risk patients into low risk (Fig. Fig. 2 Comparison of risk categorization by CAB vs Adjuvant! Online. a irrespective of node status. c Differential risk categorization by CAB vs Proportions of high- and low-risk categorization by the two tests. b Adjuvant! Online based on node status Differential risk categorization by CAB vs Adjuvant! Online S26 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Fig. 3 Physician prescriptions. a Geographical distribution of prescribers of CAB in India and outside India (n =440). b Percentage of physician prescriptions across based on number of cases prescribed Physician’s Consideration of Treatment Change Based chemotherapy, it was mostly the patient’s decision to opt for on CAB-Based Stratification chemotherapy.The prime reasons for high-risk patients who did not receive chemotherapy were older age and patient’s The goal of risk of recurrence prediction tests is to help guide preference. the physician in the decision-making of patient’s treatment planning. Thus, it was important to assess how CAB helped in guiding the treatment decision. CAB is not a stand-alone Discussion test to determine inclusion of chemotherapy into the treatment regimen and clinicians do rely on other patient-specific histor- The incidence of breast cancer in India is much lower than that ical information, comorbidities to arrive at this conclusion. in the West but nonetheless is the most common cancer in However, since low-risk patients would potentially not benefit Indian women [1]. Majority of patients present the disease at from chemotherapy while high-risk patients would benefit stages II and III with node negativity [22–24]. The median age from chemotherapy, data was analyzed for both low-risk and at onset is < 60 years, with majority of patients presenting with high-risk patients to assess if they did or did not receive che- T2 tumors. In early-stage patients, majority were diagnosed at motherapy treatment. We observed that 88% (n=254)ofpa- stage II with N0 node status [22]. More than 80% were of the tients received chemotherapy or not based on whether they invasive ductal carcinoma (IDC) morphology [23]with 20– were CAB high or low risk respectively. (Table 5). 33% hormone receptor positivity [22, 23]. The retrospective Interestingly, 92% of patients (159/173) stratified as low risk cohort used to validate CAB reflect a very similar proportions for recurrence by CAB did not receive chemotherapy as com- of clinical features with a majority of T2 tumors (65%), me- pared with 80% (65/81) of high-risk patients who received dian age at onset of < 60 years, and node negative (57%) chemotherapy. Of the low-risk patients who received indicating that the validation cohort used for CAB [15]is a true representation of the breast cancer disease characteristics reported in India. Table 5 Treatment decsions based on CAB-based risk categorization CAB is robust in risk classification and is not affected by Risk No. of Chemotherapy Chemotherapy Percentage of age at onset, luminal subtypes, or geographical locations. This category patients given not given patients could be attributed to the type of patient samples used for by CAB development and validation of the test. CAB has been shown to perform well across all age groups, irrespective of meno- Low risk 173 14 159 92% did not receive pausal status. It is also important to note that the set of bio- chemotherapy markers used in this test are not part of the ER/PR signaling or High risk 81 65 16 80% received proliferative pathways, unlike other tests like Oncotype DX chemotherapy and EndoPredict. This could explain why this test does not get Total 254 159 + 65 = 224 88% received affected by the luminal subtypes which mainly are determined chemotherapy by the expression levels of ER/PR and Ki67. On similar lines, if they were high risk and we have also shown that CAB performs better than IHC4- and no Ki67-based prognostication [15] on a retrospective cohort chemotherapy with correlation to outcome. if they were There are multiple free online tools, like Predict and low risk NPI, available that are often used by physicians as Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S27 substitutes for prognostic tests. It is however important characteristics of breast cancer in Asians (mostly to ascertain the limitations of these tools before using Chinese) are very different from those from the West them to decide patient treatment. Both NPI and Predict [28, 29]. Asians presented with much later stages of provide overall survival information and may not be disease (stage II or above), larger tumors, and a much accurate in exactly predicting benefit from chemothera- lower age at onset than their Western counterparts. It is py treatment. NPI has been shown to provide subopti- thus important to evaluate the performance of any test mal prognosis in patients who are < 40 years and sig- in the appropriate population before including them in nificantly underestimated overall survival in patients clinical practice, as noted in the ICMR guidelines for aged between 55 and 60 years [25]. Despite validation breast cancer treatment. of Predict in European, US, and Asian patients, it has CAB has been developed and validated on Indian been shown to work accurately only for Western pa- patients and have been used in clinical settings since tients within the age group of 50–65 years [26, 27]. the last 3 years. Prescriptions were higher for patients In this study, we compared the performance of with luminal B, node-negative patients where the physi- Adjuvant! Online with CAB for risk categorization. cian would want a second opinion to spare the patient We used the modified version of Adjuvant! Online from chemotherapy, who would have otherwise got it as (AOL) as described in the MINDACT study to stratify per the St. Gallen’s recommendations and other free patients into clinical low- and high-risk categories. CAB online tools. Based on the feedback from these treating categorizes more patients into low risk as compared physicians, 92% of the low-risk patients were spared with AOL as observed in both node-negative and chemotherapy. Eighty-eight percent of patients were giv- node-positive subgroups. It is interesting to note that en or not given chemotherapy depending on whether CAB identified patients who might benefit from chemo- they were CAB high or low risk respectively, indicating therapy in the clinical low-risk category with node- a reasonable physician acceptance of CAB as a test to negative disease. On the other hand, 34.4% and 47.9% effectively plan treatment decisions. Considering that AOL high-risk patients with node-positive and node- CAB is not a stand-alone test to decide treatment deci- negative disease, respectively, could avoid chemotherapy sion, it is interesting to note that we observed a higher as CAB categorized them as low risk. Based on this percentage of the low-risk group not receiving chemo- AOL and CAB comparison data, it is evident that clin- therapy (92%) than the high-risk group receiving che- ical parameters alone are not significant in predicting motherapy (80%). This indicates a paradigm shift in the risk of recurrence and the tumor biology contribution clinical management of the disease, with an intention to from the CAB biomarkers adds great value in more decrease the use of chemotherapy in treatment of early- stage patients. With all prescriptions for the test coming accurate risk prediction. St. Gallen’s 2013 guidelines suggest that only patients with in with the need to treat patients with chemotherapy, the luminal A and grade 3 disease could be given chemotherapy. change in decision from chemoendocrine therapy treat- Data from this study suggests that 17 out of the 73 (23%) who ment to endocrine therapy alone in 92% of patients is were luminal A with grade 1 or 2 stratified as high risk by very encouraging compared with the rates of 20–50% CAB would not have received chemotherapy as per these reported for other tests like ODX, MammaPrint, and guidelines leading to potential undertreatment. Two patients EndoPredict [30–34]. Though we did not observe any out of 10 who had luminal A grade 3 disease were stratified by recurrence in the low-risk patients thus far, we under- CAB as low risk and thus could avoid chemotherapy. On the stand that the 5-year follow-up period is not yet com- other hand, all luminal B patients are eligible for chemother- pleted to conclusively comment on the accuracy of pre- apy as per these guidelines. We observed that 69% (126 out of diction of recurrence by CAB. The main objective of 182) of patients with luminal B disease were low risk by CAB this study is to assess impact of CAB on physician’s and could avoid chemotherapy. Of these, 111 patients had treatment planning decision. We observed an increase node-negative disease. The data highlights the usefulness of in the adoption of the CAB by fourfold since 2016 CAB in avoiding both under- and overtreatment of patients when the test was first launched. All of these patients based on prognostic tests. It is noteworthy that the clinical tested by CAB would be followed up for 5 years to utility of prognostic tests is recognized by international guide- correlate risk prediction tooutcome and todetermine lines like AJCC (8th edition, 2018) and St. Gallen’srecom- the accuracy of risk prediction. mendations (2013) as useful tools for disease staging and to In conclusion, we observed a steady increase in the adop- decide treatment plans respectively. tion of CAB by physicians and its use for tailoring treatment Tests like MammaPrint, ODX, and EndoPredict have of patients. CAB provides affordable, accurate alternative been developed and validated on non-Asian populations. prognostic test helping up to 70% patients avoid Multiple studies have shown that the clinical chemotherapy. S28 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Acknowledgments We thank Arun Raju, Nilesh Kulkarni, Aishwariya DL, Bryant J, Wolmark N (2004) Multigene assay to predict recur- Gupta, Dr. Nirupama Naidu, and Vijaya K. for help with follow-up data rence of tamoxifen-treated, node-negative breast cancer. N Engl J collection. We thank Sukriti Malpani and Aparna Gunda for help with Med 351(27):2817–2826 data presentation. We thank all physicians who prescribed the test and 10. Van De Vijver MJ, He YD, Van’t Veer LJ et al (2002) A gene- provided us follow-up data. expression signature as a predictor of survival in breast cancer. N. Engl J Med 347(25): 1999–2009 11. Nielsen TO, Parker JS, Leung S et al (2010) A comparison of Compliance with Ethical Standards PAM50 intrinsic subtyping with immunohistochemistry and clini- cal prognostic factors in tamoxifen-treated estrogen receptor– Conflict of Interest All authors except SPS, SP, RK, KP, and DS are positive breast cancer. Clin Cancer Res 16(21):5222–5232 employees of OncoStem Diagnostics Pvt. Ltd which developed 12. Filipits M, Rudas M, Jakesz R, Dubsky P, Fitzal F, Singer CF, CanAssist-Breast. MMB is a co-inventor on a patent application related Dietze O, Greil R, Jelen A, Sevelda P, Freibauer C, Müller V, to this article. Authors do not have any other competing interest to Jänicke F, Schmidt M, Kölbl H, Rody A, Kaufmann M, Schroth declare. W, Brauch H, Schwab M, Fritz P, Weber KE, Feder IS, Hennig G, Kronenwett R, Gehrmann M, Gnant M, EP Investigators (2011) A new molecular predictor of distant recurrence in ER positive, Open Access This article is licensed under a Creative Commons HER2-negative breast cancer adds independent information to con- Attribution 4.0 International License, which permits use, sharing, adap- ventional clinical risk factors. Clin Cancer Res 17(18):6012–6020 tation, distribution and reproduction in any medium or format, as long as 13. Jerevall PL, Ma XJ, Li H, Salunga R, Kesty NC, Erlander MG, you give appropriate credit to the original author(s) and the source, pro- Sgroi DC, Holmlund B, Skoog L, Fornander T, Nordenskjöld B, vide a link to the Creative Commons licence, and indicate if changes were Stål O (2011) Prognostic utility of HOXB13: IL17BR and molec- made. The images or other third party material in this article are included ular grade index in early-stage breast cancer patients from the in the article's Creative Commons licence, unless indicated otherwise in a Stockholm trial. Br J Cancer 104(11):1762–1769 credit line to the material. If material is not included in the article's 14. Ramkumar C, Buturovic L, Malpani S et al (2018) Development of Creative Commons licence and your intended use is not permitted by a novel proteomic risk-classifier for prognostication of patients with statutory regulation or exceeds the permitted use, you will need to obtain early-stage hormone receptor–positive breast cancer. Biomark permission directly from the copyright holder. To view a copy of this Insights 13:1177271918789100 licence, visit http://creativecommons.org/licenses/by/4.0/. 15. Bakre MM, Ramkumar C, Attuluri AK et al (2019) Clinical vali- dation of an immunohistochemistry-based CanAssist-Breast test for distant recurrence prediction in hormone receptor-positive breast cancer patients. Cancer Med doi. https://doi.org/10.1002/cam4. References 16. Attuluri AK, Prakash VSC, Gunda A et al (2019) Analytical vali- dation of CanAssist-Breast: an immunohistochemistry based prog- 1. Dhillon PK, Mathur P, Nandakumar A et al (2018) The burden of nostic test for hormone receptor positive breast cancer patients. cancers and their variations across the states of India: the global BMC Cancer 19(1):249 burden of disease. Lancet Oncol 19:1289–1306 17. Goldhirsch A, Winer EP, Coates AS et al (2013) Personalizing the 2. Malvia S, Appalaraju B, Uma S et al (2017) Epidemiology of breast treatment of women with early breast cancer: highlights of the St cancer in Indian women. Asia-Pacific J Clin Oncol 13:289–295 Gallen’s International Expert consensus on the primary therapy of 3. Vostakolaei F, Karim-Kos HE, Janssen-Heijnen ML et al (2010) early breast cancer 2013. Annals Oncol 24:2206–2223 The validity of the mortality to incidence ratio as a proxy for site- 18. Cardoso F, van’t Veer LJ, Bogearts J et al (2016) 70-gene signature specific cancer survival. Eur J Pub Health 21(5):573–577 as an aid to treatment decisions in early-stage breast cancer. New 4. Sharma R (2019) Breast cancer incidence, mortality and mortality- Eng J Med 25;375(8):717-29 to-incidence ratio (MIR) are associated with human development, 19. Wishart GC, Azzato EM, Greenberg DC, Rashbass J, Kearins O, 1990–2016: evidence from Global Burden of Disease Study 2016. Lawrence G, Caldas C, Pharoah PD (2010) PREDICT: a new UK Breast Cancer doi.org/10.1007/s12282-018-00941-4 prognostic model that predicts survival following surgery for inva- 5. Wickerham LD, O’Conell MJ, Costantino JP et al (2008) The half sive breast cancer. Breast Cancer Res 12(1):R1 century of clinical trials of the National Surgical Adjuvant Breast 20. Haybittle JL, Blamey RW, Elston CW et al., A prognostic index in and Bowel Project (NSABP). Seminal Oncol 35(5):522–529 primary breast cancer. (1982) Br J Cancer 45(3):361–366 6. Schmidt M (2014) Chemotherapy in early breast cancer: when, how 21. Ravdin PM, Siminoff LA, Davis GJ, Mercer MB, Hewlett J, and which one? Breast Care 9:154–160 Gerson N, Parker HL (2001) Computer program to assist in making 7. Furquhar C, Majoribanks J, A Lethaby A et al (2016) High-dose decisions about adjuvant therapy for women with early breast can- chemotherapy and autologous bone marrow or stem cell transplan- cer. J Clin Oncol 19(4):980–991 tation versus conventional chemotherapy for women with early 22. Raina V, Bhutani M, Bedi R et al (2005) Clinical features and poor prognosis breast cancer. Cochrane Database of Systematic prognostic factors of early breast cancer at a major cancer center Reviews DOI: https://doi.org/10.1002/14651858.CD003139.pub3 in North India. Ind J Cancer 42(1): 36-41. Int Surg J 4(3): 913–915 8. Sparano JA, Gray RJ, Makower DF, Pritchard KI, Albain KS, 23. Agarwal G, Ramakant P (2008) Breast cancer care in India: current Hayes DF, Geyer CE Jr, Dees EC, Goetz MP, Olson JA Jr, Lively scenario and challenges for the future. Breast Care 3:21–27 T, Badve SS, Saphner TJ, Wagner LI, Whelan TJ, Ellis MJ, Paik S, 24. Almas A (2017) Clinical profile of carcinoma breast cases at a Wood WC, Ravdin PM, Keane MM, Gomez Moreno HL, Reddy tertiary care hospital. Int J Surg 4(3):913–915 PS, Goggins TF, Mayer IA, Brufsky AM, Toppmeyer DL, 25. Lambertini M, Pinto AC, Ameye L et al (2016) The prognostic Kaklamani VG, Berenberg JL, Abrams J, Sledge GW Jr (2018) performance of Adjuvant! Online and Nottingham Prognostic Adjuvant chemotherapy guided by a 21-gene expression assay in Index in young breast cancer patients. British J of Cancer 115: breast cancer. N Engl J Med 379(2):111–121 1471–1478 9. Paik S, Shak S, Tang G, Kim C, Baker J, Cronin M, Baehner FL, 26. Engelhardt EG, Garvelink MM, de Haes JH et al (2014) Predicting Walker MG, Watson D, Park T, Hiller W, Fisher ER, Wickerham and communicating the risk of recurrence and death in women with Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S29 early-stage breast cancer: a systematic review of risk prediction assay: real-life utilization and decision impact in French clinical models. J Clin Oncol 20 32(3):238–250 practice. BMC Cancer:39–45 32. Holt S, Bartelli G, Humphreys I et al (2013) A decision impact, 27. Maishman T, Copson E, Stanton L, Gerty S, Dicks E, Durcan L decision conflict and economic assessment of routine Oncotype DX et al. (2015) An evaluation of the prognostic model PREDICT testing of 146 women with node-negative or pNImi, ER-positive using the POSH cohort of women aged 40 years at breast cancer breast cancer in the UK B J Cancer 108, 2250–2258 diagnosis. Br J Cancer112(6):983-991 33. Müller BM, Keil E, Lehmann A, Winzer KJ, Richter-Ehrenstein C, 28. Sivasubramaniam PG, Zhang B, Zhang Q et al (2015) Breast cancer Prinzler J, Bangemann N, Reles A, Stadie S, Schoenegg W, Eucker disparities: a multicenter comparison of tumor diagnosis, character- J, Schmidt M, Lippek F, Jöhrens K, Pahl S, Sinn BV, Budczies J, istics, and surgical treatment in China and the U.S. The Oncol 20: Dietel M, Denkert C (2013) The EndoPredict gene-expression as- 1044–1050 say in clinical practice - performance and impact on clinical deci- 29. Kwang A, Cheung P, Chan S et al (2008) Breast cancer in Chinese sions. PLoS One 8(6):e68252 women younger than age 40: are they different from their solder 34. Kuijer A, Straver M, Dekker B et al (2017) Impact of 70-gene counterparts? World J Surg 32(12):2554–2561 signature use on adjuvant chemotherapy decisions in patients with 30. Lo SS, Mumby PB, Norton J et al (2010) Prospective multicenter estrogen receptor–positive early breast cancer: results of a prospec- study of the impact of the 21-gene recurrence score assay on med- tive cohort study. J Clin Oncol 24:2814–2819 ical oncologist and patient adjuvant breast cancer treatment selec- tion 10:1671-1676 31. Curtit E, Vanettzel J, Darmon J et al (2019) Results of PONDx, a Publisher’sNote Springer Nature remains neutral with regard to jurisdic- prospective multicenter study of the Oncotype DX® breast cancer tional claims in published maps and institutional affiliations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png "Indian Journal of Surgical Oncology" Springer Journals

CanAssist Breast Impacting Clinical Treatment Decisions in Early-Stage HR+ Breast Cancer Patients: Indian Scenario

Loading next page...
 
/lp/springer-journals/canassist-breast-impacting-clinical-treatment-decisions-in-early-stage-0JcLGedcq0
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2019
Subject
Medicine & Public Health; Surgical Oncology; Oncology; Surgery
ISSN
0975-7651
eISSN
0976-6952
DOI
10.1007/s13193-019-01014-4
Publisher site
See Article on Publisher Site

Abstract

CanAssist Breast (CAB) has thus far been validated on a retrospective cohort of 1123 patients who are mostly Indians. Distant metastasis–free survival (DMFS) of more than 95% was observed with significant separation (P < 0.0001) between low-risk and high-risk groups. In this study, we demonstrate the usefulness of CAB in guiding physicians to assess risk of cancer recurrence and to make informed treatment decisions for patients. Of more than 500 patients who have undergone CAB test, detailed analysis of 455 patients who were treated based on CAB-based risk predictions by more than 140 doctors across India is presented here. Majority of patients tested had node negative, T2, and grade 2 disease. Age and luminal subtypes did not affect the performance of CAB. On comparison with Adjuvant! Online (AOL), CAB categorized twice the number of patients into low risk indicating potential of overtreatment by AOL-based risk categorization. We assessed the impact of CAB testing on treatment decisions for 254 patients and observed that 92% low-risk patients were not given chemotherapy. Overall, we observed that 88% patients were either given or not given chemotherapy based on whether they were stratified as high risk or low risk for distant recurrence respectively. Based on these results, we conclude that CAB has been accepted by physicians to make treatment planning and provides a cost-effective alternative to other similar multigene prognostic tests currently available. . . . . Keywords CanAssist-Breast Decision impact Indian patients Prognostic Early-stage breast cancer Background is a composite measure of education, income, and life expec- tancy, MIR for a medium HDI country like India in 2016 was Among all cancers, breast cancer occupies the top position in found to be 0.5 compared with 0.19 for a very high HDI Indian women [1]. Based on projections, breast cancer patient country like the USA [4]. Besides economic reasons, Indian numbers could reach 1.9 million by 2020 [2]. Mortality to patients also present with more aggressive form of the disease, incidence rate (MIR) is a useful indicator of 5-year survival with an earlier age at onset and higher tumor stage and node across various cancers [3]. In a study stratifying MIR for positivity. Even with the presence of high risk factors, not breast cancer with Human Development Index (HDI) which every patient might benefit from chemotherapy and the chal- lenge is in identifying these patients accurately. In the past 60 years, since the first cancer patient enrolled * Manjiri M. Bakre in a randomized clinical trial conducted by the National manjiri@oncostemdiagnostics.com Surgical Adjuvant Breast and Bowel Project (NSABP) [5], there has been constant improvement in the effective OncoStem Diagnostics Private Limited, 4, Raja Ram Mohan Roy Road, Aanand Towers, 2nd Floor, Bangalore, Karnataka 560027, treatment and surgical management of breast cancer patients. India Starting with a combination of cyclophosphamide, Manipal Hospital, Bangalore, India methotrexate, and 5-flurouracil (CMF) in the 1950s, several 3 new drugs (anthracyclines and taxanes) and drug regimens Sri Shankara Cancer Hospital and Research Center, Bangalore, India have since evolved [6], in addition to hormone therapy, HCG, Bangalore, India for hormone receptor–positive breast cancer. While the Rajiv Gandhi Cancer Institute and Research Center, New Delhi, India efficacy improved in some cases, these drugs also produced Mangalore Institute of Oncology, Manipal, India unwanted side effects like cardiotoxicity and decrease in Father Muller Medical College, Mangalore, India blood cell counts. S22 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Optimal treatment helps in minimizing mortality and The Indian Council of Medical Research (ICMR) does not morbidity associated with the disease. Treating every include tests like Oncotype DX and MammaPrint in their rec- patient with the most aggressive form of treatment ommendations for treatment decision-making due to the lack may not always be productive, even if the patient does of data on Indian patients (https://www.icmr.nic.in/sites/ not have existing comorbidities. The Cochrane review default/files/guidelines/Breast_Cancer.pdf). The critical point compiled data from 14 randomized clinical trials involv- of development and validation on Indian patients makes ing 5600 women. There was quality data to show that CanAssist-Breast a good choice for Indian patients. high doses of chemotherapy did not improve survival in CanAssist Breast test has been used for treatment planning early-stage breast cancer patients [7]. The harmful side for patients since 2016. The aims of the present study are to effects of chemotherapy are well known and balancing briefly summarize the retrospective validation data about the harm to benefit from chemotherapy would improve CAB and analyze the clinical scenarios for recommending this the quality of life of patients. With the advent of test and correlation of risk stratification by CAB to treatment multigene prognostic tests, studies have shown 70% decisions by referring physicians. early-stage hormone receptor–positive, node-negative breast cancer patients benefit from chemotherapy [8]. Some of the commercially available prognostic tests in- Methods clude Oncotype DX (ODX) [9], MammaPrint [10], Prosigna [11], EndoPredict [12], Breast Cancer Index (BCI) [13], and Patient Selection CanAssist Breast (CAB) [14]. All these tests query different gene/protein markers and use different testing methodologies For the retrospective study, we obtained post-surgical tumor (Table 1) but have been shown to predict risk of recurrence samples in the form of formalin-fixed paraffin-embedded with more than 95% accuracy. (FFPE) blocks from 1123 early-stage breast cancer patients. Of all these tests, CAB is the only test that has been Patient consent, patient information, and treatment follow-up developed and validated on a mixed cohort of Asian details such as age, year of diagnosis, type of surgery, tumor (Indian) and Caucasian patients in a 3:1 proportion size and grade, hormone receptor status, node status, treatment [15] and extensively validated for analytical perfor- regimen, date of recurrence or last visit, or death were obtain- mance [16]. CAB is an immunohistochemistry (IHC)- ed from the treating hospitals. All patients had hormone based test that assesses risk of cancer recurrence at a receptor–positive disease. TNBC patients were excluded from distant site within 5 years from diagnosis. It quantifies the study. The patients were staged based on the AJCC 7th protein expression levels of a combination of 5 unique edition staging system. Patients with tumors with stage I non-proliferative biomarkers (CD44, Pan-Cadherin, N- (T1N0) and stage II (T1N1, T2N0, T2N1, T3N0) were con- Cadherin, ABCC4, and ABCC11). CAB markers are sidered as early stage. Patients with a minimum of 5-year involved in diverse cancer signaling pathways that reg- follow-up were included and this requirement was waived ulate cancer metastasis and drug resistance. The IHC off only in patients with a recurrence at a distant site within data from the biomarkers are combined with three clin- the 5-year period. ical parameters, tumor size (T), node status (N), and For the decision impact study, patients who were recom- tumor grade, to generate a low- or a high-risk score mended testing by CanAssist Breast by their treating physi- for every patient using a machine learning–based statis- cian between 2016 and May 2019 were included in this study. tical model. Patient consent was taken as part of test requisition form. Table 1 Summary of commercially available prognostic tests Oncotype DX MammaPrint Prosigna EndoPredict Breast CanAssist Breast Cancer Index No. of genes used 21 70 50 8 7 5 biomarkers Proliferation genes included Yes Yes Yes Yes No No Method qPCR DNA microarray NanoString qPCR qPCR IHC Clinical parameters No No No Tumor size, No Tumor size, node status node status, and tumor grade Presence of intermediate zone Yes No Yes No No No Prediction of chemotherapy benefit Yes No No No No Yes Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S23 Clinical information like node, tumor size, and tumor grade CAP-accredited central reference OncoStem lab at Bangalore, was obtained from pathology reports obtained from the pa- India. The IHC grading information was incorporated along tients along with the test requisition form. ER/PR status was with the 3 clinical parameters (tumor size, tumor grade, and included as a mandatory requirement for acceptance for nodal status) as obtained from the treating hospitals to calcu- CanAssist Breast testing and the report was provided by the late CAB risk scores on a scale of 0–100. A cutoff of 15.5 is patient or hospital from where the cases were referred by. used to classify patients into low risk (score ≤ 15.5) or high Ki67, wherever available along with the ER/PR reports, was risk (score > 15.5) for distant recurrence [14]. used for luminal classification in this study. Referring physi- cian name and patient age were obtained from the test requi- sition form. Luminal subtyping was performed as per St. Adjuvant! Online–Based Risk Categorization Gallen’s recommendations [17]. The modified Adjuvant! Online (ver8) criteria as described in Study Design the MINDACT trial [18] was used for assigning risk catego- ries based on tumor grade, node status, and tumor size A total of 455 patients who were recommended CAB for (Table 2). A total of 430 cases for which the exact T size and number of metastatic nodes data were available were used treatment planning were included for various subgroup anal- yses in this study. To assess region-wise distribution of pre- in this concordance analysis. scribing physicians from India, five different zones were con- sidered. These included North (Delhi, Uttar Pradesh, Haryana, and Rajasthan), Central (Madhya Pradesh), West Results (Maharashtra, Gujarat, and Rajasthan), East (West Begal), and South (Karnataka, Kerala, Andhra Pradesh, and Tamil Performance of CAB on Retrospective Cohort Nadu). Prescriptions were also received from neighboring countries of the Indian subcontinent comprising of Sri Retrospective patients (n = 1123) were dichotomized into Lanka, Pakistan, and Bangladesh which were considered as “outside India” for analysis. Information on usefulness of low- and high-risk groups by CAB. There was 11% difference in the DMFS between low- and high-risk groups demonstrat- CAB-based risk stratification for treatment planning was ob- tained from prescribing physicians via email and personal ing statistically significant (P < 0.0001) separation between the two groups (Fig. 1a). Since this cohort had a mix of pa- visits. Data was collected and analyzed to assess if patients tients with and without chemotherapy treatment, we further stratified as either low risk or high risk for recurrence by CAB performed survival analysis on patients who were treated with were given chemotherapy or not for patients who were pre- endocrine therapy alone (n = 298). This was to exclude any scribed CAB. This follow-up information was obtained for 254 patients. confounding effect of chemotherapy in the mixed cohort. The separation between the low- and high-risk groups was signif- Statistical Analysis icant (P = 0.0002) for this subgroup as well (Fig. 1b)with a clear difference in DMFS of 14% between the two risk groups. Kaplan-Meier survival curve analysis was performed using GraphPad version 8. Distant metastasis–free survival Table 2 Criteria used for clinical risk classification (DMFS) was calculated for CAB high-risk versus low-risk patients. P values were computed using log-rank two-sided Grade Nodal status Tumor size Clinical risk test at 0.05 significance. DMFS is the time interval between Well differentiated N0 ≤3cm C-Low the date of diagnosis of cancer and the last date of follow-up in (grade 1) 3.1–5cm C-High case of no event/recurrence with a minimum period of 5 years. N1 (1–3 ≤2cm C-Low MedCalc (https://www.medcalc.org/calc/comparison_of_ positive nodes) 2.1–5cm C-High proportions.php) was used to calculate statistical significance Moderately differentiated N0 ≤2cm C-Low of proportions. P values less than or equal to 0.05 were (grade 2) 2.1–5cm C-High considered to be statistically significant. N1 (1–3 Any size C-High positive nodes) CanAssist Breast Testing Poorly differentiated N0 ≤1cm C-Low (grade 3) 1.1–5cm C-High Testing was performed on FFPE blocks. The CAB was per- N1 (1–3 Any size C-High formed as described previously [14, 15]. Briefly, immunohis- positive nodes) tochemistry was performed for the 5 CAB biomarkers at the S24 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Fig. 1 Performance of CAB. a Survival (KM) curve using CAB- based risk categorization on a retrospective mixed cohort of chemotherapy-treated and chemotherapy-naïve (endocrine therapy alone treated) patients. b Survival analysis using CAB- based categorization with chemotherapy-naïve patient cohort The DMFS for both the mixed and chemotherapy-naïve sub- clinical characteristics were predominantly similar to that of groups was > 95% in the low-risk group. the retrospective cohort used for test validation. In the node-positive group of the prospective cohort, 94% of patients presented with N1 disease. Sixty-five percent of the Evaluation of Various Risk Factors in the Study Cohort patients with N1 disease were single-node positive, 16% were in the Prospective Cohort double-node positive, and 13% were triple-node positive. Based on the tumor sizes, 92% of T1 patients had a tumor Upon successful completion of clinical, analytical validation of size of ≥ 1 cm; 85% and 27% of T2 patients had tumor sizes CanAssist-Breast and upon getting the appropriate regulatory between 2–3and 3–4 cm respectively; 1.8% of patients had accreditations, CAB was launched in the market from the mid- tumor size > 5 cm. Percentage of grades 1 and 3 were 19.8 and dle of 2016. Until date, over 500 patients have availed CAB in 14.9% respectively. the Indian subcontinent. Analysis has been shown in the sub- 55.4% of patients were in the age group of 41–60 years and sequent sections on the prospective cohort of 455 patients. 5% patients had an age at onset of ≤ 40 years. 68.7% of pa- As shown in the retrospective clinical validation cohort data tients were of the luminal B subtype (Table 3). [15], we had observed that most patients had T2 tumors with grade 2 disease and node-negative disease [15]. We wanted to evaluate if this was the case with prospective patient cohort Clinical Factors Influencing CAB Risk Categorization who were prescribed CAB. We therefore performed subgroup analysis based on node status, tumor size, tumor grade, and We further evaluated the effect of the clinical factors on CAB- based risk categorization. Luminal subtype and age had no age. There was a high proportion of patients with N0 (82.4%), T2 (58.4%), and grade 2 (65.3%) disease (Table 3). Thus, significant effect on proportions of low-risk stratified patients by CAB, while the influence of tumor size, node, and tumor grade was significant (Table 4). We also evaluated the perfor- Table 3 Cohorts’ description (n = 455). For T size, n =448; luminal mance of the test across patients from two geographically subtyping, 265 patients for whom Ki67 status was known were distinct zones (North and South) of India and found that the considered proportions of low- and high-risk patients was not Parameter Number of patients (%) T1 178 (39.6) Table 4 Significance of low-risk proportions across various disease T2 262 (58.4) parameters T3 9 (2.0) Parameter % low risk P value N0 375 (82.4) N+ 80 (17.6) Luminal A 70 P =0.5 Luminal B 69 G1 90 (19.8) T1 83 P <0.0001 G2 297 (65.3) T2+T3 61 G3 68 (14.9) N0 74 P =0.0004 Luminal A 83 (31.3) N+ 54 Luminal B 182 (68.7) G1 89 P <0.0001 Age < 40 years 23 (5.0) G2+G3 65 Age 41–60 years 252 (55.4) <40 years 74 P =0.7 Age > 61 years 180 (39.6) >40 years 70 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S25 significantly different across these two patient subgroups (P = 2b). CAB categorized 34.4% N0 and 47.9% of N+ 0.8), and to the total pooled subgroup (P = 0.7 for South vs AOL (clinical) high-risk patients into low risk (Fig. 2c). total; P = 0.6 for North vs total). Comparison of CAB with Adjuvant! Online on Risk Physician Considerations for Testing Categorization The tests were prescribed by 147 physicians from all across In the scenario where patients cannot afford expensive India (North, South, Central, East, and West) and from neigh- prognostic tests, physicians tend to decide on therapy boring countries (Fig. 3a). While 47% of physicians pre- options based on various clinical parameters and freely scribed the test for more than one patient with 11% of physi- cians prescribing > 6 patients for CAB testing (Fig. 3b). Next, available online predictive tools like Predict [19], NPI [20], and Adjuvant! Online (AOL) [21]. While NPI and we wanted to assess the reasons for prescribing CAB by phy- sicians to their patients. From the survey conducted by us, the Predict provide estimate of overall survival, Adjuvant! Online provides risk of recurrence and potential re- major reasons were (a) presentation with high-risk clinical features like large tumors or node positivity, (b) to assess if sponse to chemotherapy. Since both CAB and Adjuvant! Online aid in making treatment decisions, patients with luminal A disease would benefit from chemo- therapy, (c) young age of patients who want to avoid side we compared the performance of CAB with AOL that predicts risk based on clinical parameters alone. For effects of chemotherapy, (d) old aged patients who might be those patients for whom we had information of the ex- spared chemotherapy if possible, and (e) comorbidities in the act size of the tumor (n = 430), we compared the risk patient that might require evaluating the benefit of chemother- categorization by AOL vs CAB. AOL (clinical) catego- apy over its effect that would have a bearing on the quality of life. Across all these reasons, the main rationale for prescrip- rized 62.4% patients into high risk as compared with 29% by CAB (Fig. 2a). CAB categorized 36.6% of tion for all the physicians was to avoid chemotherapy, if possible. AOL (clinical) high-risk patients into low risk (Fig. Fig. 2 Comparison of risk categorization by CAB vs Adjuvant! Online. a irrespective of node status. c Differential risk categorization by CAB vs Proportions of high- and low-risk categorization by the two tests. b Adjuvant! Online based on node status Differential risk categorization by CAB vs Adjuvant! Online S26 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Fig. 3 Physician prescriptions. a Geographical distribution of prescribers of CAB in India and outside India (n =440). b Percentage of physician prescriptions across based on number of cases prescribed Physician’s Consideration of Treatment Change Based chemotherapy, it was mostly the patient’s decision to opt for on CAB-Based Stratification chemotherapy.The prime reasons for high-risk patients who did not receive chemotherapy were older age and patient’s The goal of risk of recurrence prediction tests is to help guide preference. the physician in the decision-making of patient’s treatment planning. Thus, it was important to assess how CAB helped in guiding the treatment decision. CAB is not a stand-alone Discussion test to determine inclusion of chemotherapy into the treatment regimen and clinicians do rely on other patient-specific histor- The incidence of breast cancer in India is much lower than that ical information, comorbidities to arrive at this conclusion. in the West but nonetheless is the most common cancer in However, since low-risk patients would potentially not benefit Indian women [1]. Majority of patients present the disease at from chemotherapy while high-risk patients would benefit stages II and III with node negativity [22–24]. The median age from chemotherapy, data was analyzed for both low-risk and at onset is < 60 years, with majority of patients presenting with high-risk patients to assess if they did or did not receive che- T2 tumors. In early-stage patients, majority were diagnosed at motherapy treatment. We observed that 88% (n=254)ofpa- stage II with N0 node status [22]. More than 80% were of the tients received chemotherapy or not based on whether they invasive ductal carcinoma (IDC) morphology [23]with 20– were CAB high or low risk respectively. (Table 5). 33% hormone receptor positivity [22, 23]. The retrospective Interestingly, 92% of patients (159/173) stratified as low risk cohort used to validate CAB reflect a very similar proportions for recurrence by CAB did not receive chemotherapy as com- of clinical features with a majority of T2 tumors (65%), me- pared with 80% (65/81) of high-risk patients who received dian age at onset of < 60 years, and node negative (57%) chemotherapy. Of the low-risk patients who received indicating that the validation cohort used for CAB [15]is a true representation of the breast cancer disease characteristics reported in India. Table 5 Treatment decsions based on CAB-based risk categorization CAB is robust in risk classification and is not affected by Risk No. of Chemotherapy Chemotherapy Percentage of age at onset, luminal subtypes, or geographical locations. This category patients given not given patients could be attributed to the type of patient samples used for by CAB development and validation of the test. CAB has been shown to perform well across all age groups, irrespective of meno- Low risk 173 14 159 92% did not receive pausal status. It is also important to note that the set of bio- chemotherapy markers used in this test are not part of the ER/PR signaling or High risk 81 65 16 80% received proliferative pathways, unlike other tests like Oncotype DX chemotherapy and EndoPredict. This could explain why this test does not get Total 254 159 + 65 = 224 88% received affected by the luminal subtypes which mainly are determined chemotherapy by the expression levels of ER/PR and Ki67. On similar lines, if they were high risk and we have also shown that CAB performs better than IHC4- and no Ki67-based prognostication [15] on a retrospective cohort chemotherapy with correlation to outcome. if they were There are multiple free online tools, like Predict and low risk NPI, available that are often used by physicians as Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S27 substitutes for prognostic tests. It is however important characteristics of breast cancer in Asians (mostly to ascertain the limitations of these tools before using Chinese) are very different from those from the West them to decide patient treatment. Both NPI and Predict [28, 29]. Asians presented with much later stages of provide overall survival information and may not be disease (stage II or above), larger tumors, and a much accurate in exactly predicting benefit from chemothera- lower age at onset than their Western counterparts. It is py treatment. NPI has been shown to provide subopti- thus important to evaluate the performance of any test mal prognosis in patients who are < 40 years and sig- in the appropriate population before including them in nificantly underestimated overall survival in patients clinical practice, as noted in the ICMR guidelines for aged between 55 and 60 years [25]. Despite validation breast cancer treatment. of Predict in European, US, and Asian patients, it has CAB has been developed and validated on Indian been shown to work accurately only for Western pa- patients and have been used in clinical settings since tients within the age group of 50–65 years [26, 27]. the last 3 years. Prescriptions were higher for patients In this study, we compared the performance of with luminal B, node-negative patients where the physi- Adjuvant! Online with CAB for risk categorization. cian would want a second opinion to spare the patient We used the modified version of Adjuvant! Online from chemotherapy, who would have otherwise got it as (AOL) as described in the MINDACT study to stratify per the St. Gallen’s recommendations and other free patients into clinical low- and high-risk categories. CAB online tools. Based on the feedback from these treating categorizes more patients into low risk as compared physicians, 92% of the low-risk patients were spared with AOL as observed in both node-negative and chemotherapy. Eighty-eight percent of patients were giv- node-positive subgroups. It is interesting to note that en or not given chemotherapy depending on whether CAB identified patients who might benefit from chemo- they were CAB high or low risk respectively, indicating therapy in the clinical low-risk category with node- a reasonable physician acceptance of CAB as a test to negative disease. On the other hand, 34.4% and 47.9% effectively plan treatment decisions. Considering that AOL high-risk patients with node-positive and node- CAB is not a stand-alone test to decide treatment deci- negative disease, respectively, could avoid chemotherapy sion, it is interesting to note that we observed a higher as CAB categorized them as low risk. Based on this percentage of the low-risk group not receiving chemo- AOL and CAB comparison data, it is evident that clin- therapy (92%) than the high-risk group receiving che- ical parameters alone are not significant in predicting motherapy (80%). This indicates a paradigm shift in the risk of recurrence and the tumor biology contribution clinical management of the disease, with an intention to from the CAB biomarkers adds great value in more decrease the use of chemotherapy in treatment of early- stage patients. With all prescriptions for the test coming accurate risk prediction. St. Gallen’s 2013 guidelines suggest that only patients with in with the need to treat patients with chemotherapy, the luminal A and grade 3 disease could be given chemotherapy. change in decision from chemoendocrine therapy treat- Data from this study suggests that 17 out of the 73 (23%) who ment to endocrine therapy alone in 92% of patients is were luminal A with grade 1 or 2 stratified as high risk by very encouraging compared with the rates of 20–50% CAB would not have received chemotherapy as per these reported for other tests like ODX, MammaPrint, and guidelines leading to potential undertreatment. Two patients EndoPredict [30–34]. Though we did not observe any out of 10 who had luminal A grade 3 disease were stratified by recurrence in the low-risk patients thus far, we under- CAB as low risk and thus could avoid chemotherapy. On the stand that the 5-year follow-up period is not yet com- other hand, all luminal B patients are eligible for chemother- pleted to conclusively comment on the accuracy of pre- apy as per these guidelines. We observed that 69% (126 out of diction of recurrence by CAB. The main objective of 182) of patients with luminal B disease were low risk by CAB this study is to assess impact of CAB on physician’s and could avoid chemotherapy. Of these, 111 patients had treatment planning decision. We observed an increase node-negative disease. The data highlights the usefulness of in the adoption of the CAB by fourfold since 2016 CAB in avoiding both under- and overtreatment of patients when the test was first launched. All of these patients based on prognostic tests. It is noteworthy that the clinical tested by CAB would be followed up for 5 years to utility of prognostic tests is recognized by international guide- correlate risk prediction tooutcome and todetermine lines like AJCC (8th edition, 2018) and St. Gallen’srecom- the accuracy of risk prediction. mendations (2013) as useful tools for disease staging and to In conclusion, we observed a steady increase in the adop- decide treatment plans respectively. tion of CAB by physicians and its use for tailoring treatment Tests like MammaPrint, ODX, and EndoPredict have of patients. CAB provides affordable, accurate alternative been developed and validated on non-Asian populations. prognostic test helping up to 70% patients avoid Multiple studies have shown that the clinical chemotherapy. S28 Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 Acknowledgments We thank Arun Raju, Nilesh Kulkarni, Aishwariya DL, Bryant J, Wolmark N (2004) Multigene assay to predict recur- Gupta, Dr. Nirupama Naidu, and Vijaya K. for help with follow-up data rence of tamoxifen-treated, node-negative breast cancer. N Engl J collection. We thank Sukriti Malpani and Aparna Gunda for help with Med 351(27):2817–2826 data presentation. We thank all physicians who prescribed the test and 10. Van De Vijver MJ, He YD, Van’t Veer LJ et al (2002) A gene- provided us follow-up data. expression signature as a predictor of survival in breast cancer. N. Engl J Med 347(25): 1999–2009 11. Nielsen TO, Parker JS, Leung S et al (2010) A comparison of Compliance with Ethical Standards PAM50 intrinsic subtyping with immunohistochemistry and clini- cal prognostic factors in tamoxifen-treated estrogen receptor– Conflict of Interest All authors except SPS, SP, RK, KP, and DS are positive breast cancer. Clin Cancer Res 16(21):5222–5232 employees of OncoStem Diagnostics Pvt. Ltd which developed 12. Filipits M, Rudas M, Jakesz R, Dubsky P, Fitzal F, Singer CF, CanAssist-Breast. MMB is a co-inventor on a patent application related Dietze O, Greil R, Jelen A, Sevelda P, Freibauer C, Müller V, to this article. Authors do not have any other competing interest to Jänicke F, Schmidt M, Kölbl H, Rody A, Kaufmann M, Schroth declare. W, Brauch H, Schwab M, Fritz P, Weber KE, Feder IS, Hennig G, Kronenwett R, Gehrmann M, Gnant M, EP Investigators (2011) A new molecular predictor of distant recurrence in ER positive, Open Access This article is licensed under a Creative Commons HER2-negative breast cancer adds independent information to con- Attribution 4.0 International License, which permits use, sharing, adap- ventional clinical risk factors. Clin Cancer Res 17(18):6012–6020 tation, distribution and reproduction in any medium or format, as long as 13. Jerevall PL, Ma XJ, Li H, Salunga R, Kesty NC, Erlander MG, you give appropriate credit to the original author(s) and the source, pro- Sgroi DC, Holmlund B, Skoog L, Fornander T, Nordenskjöld B, vide a link to the Creative Commons licence, and indicate if changes were Stål O (2011) Prognostic utility of HOXB13: IL17BR and molec- made. The images or other third party material in this article are included ular grade index in early-stage breast cancer patients from the in the article's Creative Commons licence, unless indicated otherwise in a Stockholm trial. Br J Cancer 104(11):1762–1769 credit line to the material. If material is not included in the article's 14. Ramkumar C, Buturovic L, Malpani S et al (2018) Development of Creative Commons licence and your intended use is not permitted by a novel proteomic risk-classifier for prognostication of patients with statutory regulation or exceeds the permitted use, you will need to obtain early-stage hormone receptor–positive breast cancer. Biomark permission directly from the copyright holder. To view a copy of this Insights 13:1177271918789100 licence, visit http://creativecommons.org/licenses/by/4.0/. 15. Bakre MM, Ramkumar C, Attuluri AK et al (2019) Clinical vali- dation of an immunohistochemistry-based CanAssist-Breast test for distant recurrence prediction in hormone receptor-positive breast cancer patients. Cancer Med doi. https://doi.org/10.1002/cam4. References 16. Attuluri AK, Prakash VSC, Gunda A et al (2019) Analytical vali- dation of CanAssist-Breast: an immunohistochemistry based prog- 1. Dhillon PK, Mathur P, Nandakumar A et al (2018) The burden of nostic test for hormone receptor positive breast cancer patients. cancers and their variations across the states of India: the global BMC Cancer 19(1):249 burden of disease. Lancet Oncol 19:1289–1306 17. Goldhirsch A, Winer EP, Coates AS et al (2013) Personalizing the 2. Malvia S, Appalaraju B, Uma S et al (2017) Epidemiology of breast treatment of women with early breast cancer: highlights of the St cancer in Indian women. Asia-Pacific J Clin Oncol 13:289–295 Gallen’s International Expert consensus on the primary therapy of 3. Vostakolaei F, Karim-Kos HE, Janssen-Heijnen ML et al (2010) early breast cancer 2013. Annals Oncol 24:2206–2223 The validity of the mortality to incidence ratio as a proxy for site- 18. Cardoso F, van’t Veer LJ, Bogearts J et al (2016) 70-gene signature specific cancer survival. Eur J Pub Health 21(5):573–577 as an aid to treatment decisions in early-stage breast cancer. New 4. Sharma R (2019) Breast cancer incidence, mortality and mortality- Eng J Med 25;375(8):717-29 to-incidence ratio (MIR) are associated with human development, 19. Wishart GC, Azzato EM, Greenberg DC, Rashbass J, Kearins O, 1990–2016: evidence from Global Burden of Disease Study 2016. Lawrence G, Caldas C, Pharoah PD (2010) PREDICT: a new UK Breast Cancer doi.org/10.1007/s12282-018-00941-4 prognostic model that predicts survival following surgery for inva- 5. Wickerham LD, O’Conell MJ, Costantino JP et al (2008) The half sive breast cancer. Breast Cancer Res 12(1):R1 century of clinical trials of the National Surgical Adjuvant Breast 20. Haybittle JL, Blamey RW, Elston CW et al., A prognostic index in and Bowel Project (NSABP). Seminal Oncol 35(5):522–529 primary breast cancer. (1982) Br J Cancer 45(3):361–366 6. Schmidt M (2014) Chemotherapy in early breast cancer: when, how 21. Ravdin PM, Siminoff LA, Davis GJ, Mercer MB, Hewlett J, and which one? Breast Care 9:154–160 Gerson N, Parker HL (2001) Computer program to assist in making 7. Furquhar C, Majoribanks J, A Lethaby A et al (2016) High-dose decisions about adjuvant therapy for women with early breast can- chemotherapy and autologous bone marrow or stem cell transplan- cer. J Clin Oncol 19(4):980–991 tation versus conventional chemotherapy for women with early 22. Raina V, Bhutani M, Bedi R et al (2005) Clinical features and poor prognosis breast cancer. Cochrane Database of Systematic prognostic factors of early breast cancer at a major cancer center Reviews DOI: https://doi.org/10.1002/14651858.CD003139.pub3 in North India. Ind J Cancer 42(1): 36-41. Int Surg J 4(3): 913–915 8. Sparano JA, Gray RJ, Makower DF, Pritchard KI, Albain KS, 23. Agarwal G, Ramakant P (2008) Breast cancer care in India: current Hayes DF, Geyer CE Jr, Dees EC, Goetz MP, Olson JA Jr, Lively scenario and challenges for the future. Breast Care 3:21–27 T, Badve SS, Saphner TJ, Wagner LI, Whelan TJ, Ellis MJ, Paik S, 24. Almas A (2017) Clinical profile of carcinoma breast cases at a Wood WC, Ravdin PM, Keane MM, Gomez Moreno HL, Reddy tertiary care hospital. Int J Surg 4(3):913–915 PS, Goggins TF, Mayer IA, Brufsky AM, Toppmeyer DL, 25. Lambertini M, Pinto AC, Ameye L et al (2016) The prognostic Kaklamani VG, Berenberg JL, Abrams J, Sledge GW Jr (2018) performance of Adjuvant! Online and Nottingham Prognostic Adjuvant chemotherapy guided by a 21-gene expression assay in Index in young breast cancer patients. British J of Cancer 115: breast cancer. N Engl J Med 379(2):111–121 1471–1478 9. Paik S, Shak S, Tang G, Kim C, Baker J, Cronin M, Baehner FL, 26. Engelhardt EG, Garvelink MM, de Haes JH et al (2014) Predicting Walker MG, Watson D, Park T, Hiller W, Fisher ER, Wickerham and communicating the risk of recurrence and death in women with Indian J Surg Oncol (April 2021) 12 (Suppl 1):S21–S29 S29 early-stage breast cancer: a systematic review of risk prediction assay: real-life utilization and decision impact in French clinical models. J Clin Oncol 20 32(3):238–250 practice. BMC Cancer:39–45 32. Holt S, Bartelli G, Humphreys I et al (2013) A decision impact, 27. Maishman T, Copson E, Stanton L, Gerty S, Dicks E, Durcan L decision conflict and economic assessment of routine Oncotype DX et al. (2015) An evaluation of the prognostic model PREDICT testing of 146 women with node-negative or pNImi, ER-positive using the POSH cohort of women aged 40 years at breast cancer breast cancer in the UK B J Cancer 108, 2250–2258 diagnosis. Br J Cancer112(6):983-991 33. Müller BM, Keil E, Lehmann A, Winzer KJ, Richter-Ehrenstein C, 28. Sivasubramaniam PG, Zhang B, Zhang Q et al (2015) Breast cancer Prinzler J, Bangemann N, Reles A, Stadie S, Schoenegg W, Eucker disparities: a multicenter comparison of tumor diagnosis, character- J, Schmidt M, Lippek F, Jöhrens K, Pahl S, Sinn BV, Budczies J, istics, and surgical treatment in China and the U.S. The Oncol 20: Dietel M, Denkert C (2013) The EndoPredict gene-expression as- 1044–1050 say in clinical practice - performance and impact on clinical deci- 29. Kwang A, Cheung P, Chan S et al (2008) Breast cancer in Chinese sions. PLoS One 8(6):e68252 women younger than age 40: are they different from their solder 34. Kuijer A, Straver M, Dekker B et al (2017) Impact of 70-gene counterparts? World J Surg 32(12):2554–2561 signature use on adjuvant chemotherapy decisions in patients with 30. Lo SS, Mumby PB, Norton J et al (2010) Prospective multicenter estrogen receptor–positive early breast cancer: results of a prospec- study of the impact of the 21-gene recurrence score assay on med- tive cohort study. J Clin Oncol 24:2814–2819 ical oncologist and patient adjuvant breast cancer treatment selec- tion 10:1671-1676 31. Curtit E, Vanettzel J, Darmon J et al (2019) Results of PONDx, a Publisher’sNote Springer Nature remains neutral with regard to jurisdic- prospective multicenter study of the Oncotype DX® breast cancer tional claims in published maps and institutional affiliations.

Journal

"Indian Journal of Surgical Oncology"Springer Journals

Published: Dec 9, 2019

References