Predicting Pathologic Complete Response after Neoadjuvant Chemotherapy

HomeRadiologyVol. 299, No. 2 PreviousNext Reviews and CommentaryFree AccessEditorialPredicting Pathologic Complete Response after Neoadjuvant ChemotherapyMassimo Imbriaco , Andrea PonsiglioneMassimo Imbriaco , Andrea PonsiglioneAuthor AffiliationsFrom the Department of Advanced Biomedical Sciences, University of Naples Federico II, Via Pansini 5, 80131 Naples, Italy.Address correspondence to M.I. (e-mail: [email protected]).Massimo Imbriaco Andrea PonsiglionePublished Online:Mar 23 2021https://doi.org/10.1148/radiol.2021210138MoreSectionsPDF ToolsImage ViewerAdd to favoritesCiteTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked In See also the article by Kim et al in this issue.Massimo Imbriaco, MD, was board certified in nuclear medicine and radiology at the University of Naples Federico II in Italy and in 2015 joined as Associate Professor of Radiology, Department of Advanced Biomedical Sciences. Dr Imbriaco has long-standing experience in body imaging with emphasis on cardiovascular, breast, and prostatic imaging; has lectured extensively both within Europe and internationally; and has authored over 200 peer-reviewed articles.Download as PowerPointOpen in Image Viewer Andrea Ponsiglione, MD, was board certified in radiology at the University of Naples Federico II in Italy. His primary research interest is in body MR and CT imaging, and he has authored over 30 peer-reviewed articles. Dr Ponsiglione is currently a PhD student at the University of Naples Federico II.Download as PowerPointOpen in Image Viewer Neoadjuvant chemotherapy (NAC) is a powerful and well-established treatment strategy that is largely used in patients with locally advanced breast cancer. The primary goal of NAC is to reduce tumor size and axillary lymph node burden, allowing pathologic complete response (pCR) and, thus, improving disease-free survival (1). In addition, NAC can facilitate sufficient tumor shrinkage to enable breast-conserving surgery in select patients without substantial increases in local recurrence and has shown some promise as a prognostic marker for patient outcome (2).As Kim et al (3) observe in this issue of Radiology, the rates of pCR differ according to tumor subtype. In particular, estrogen receptor negativity, low T stage, high histologic grade, and high Ki-67 proliferation index have been associated with pCR (4). Several studies have attempted to develop nomograms to predict pCR mainly based on the extent of expression of biologic characteristics of the primary tumor as well as the duration of preoperative chemotherapy (4).The accurate identification of factors predictive of response to preoperative treatment programs continues to represent a major research challenge. Moreover, intensive studies aiming to find clinical or molecular biomarkers that can predict the efficacy of NAC and identify patients who can benefit from chemotherapy are currently ongoing.Among imaging modalities, dynamic contrast-enhanced breast MRI, because of its high sensitivity, is widely used to assess the extent of disease before treatment and the response of the tumor to NAC. Dynamic contrast-enhanced MRI showed the highest diagnostic accuracy when compared with clinical examination and other conventional imaging modalities (eg, mammography and US) (5). More recently, several studies have focused on detecting breast cancer subtypes or characterizing and predicting pCR after NAC, using radiomics features reported on dynamic contrast-enhanced MRI scans, apparent diffusion coefficient maps, or both (6).In their single-center retrospective work, Kim et al (3) propose a comprehensive nomogram based on a combination of MRI and clinical-pathologic variables to predict pCR in patients with stage II–III breast cancer.The study is a continuation of two previous works, and there is patient overlap (7,8). In particular, of the 710 patients in the current study, 60 were reported in the first trial (7) and 154 in the second (8). The prior studies focused on the correlations of pathologic and imaging parameters, whereas the current work by Kim et al expands on these studies by including a significantly larger patient population and proposing and validating an innovative and original nomogram through the integration of clinical-pathologic parameters with MRI findings. The final study cohort included a development cohort (n = 359) and a validation cohort (n = 351). Clinical-pathologic data were collected, and mammograms and MRI scans obtained before and after NAC were analyzed.A major finding of the study by Kim et al is that, when only considering the clinical-pathologic model—which included low T stage (clinical T1 or T2), hormone receptor negativity, and high Ki-67 index—the area under the receiver operating characteristic curve (AUC) was 0.82, which is in agreement with the ranges of previously published nomograms (range, 0.76–0.83) (4,9). Conversely, in the combined model incorporating both clinical-pathologic and imaging variables, hormone receptor negativity, high Ki-67 index, and post-NAC MRI variables—including small tumor size, low lesion-to-background parenchymal signal enhancement ratio, and absence of enhancement in the tumor bed—were independently associated with pCR, with an AUC of 0.92, which is significantly higher than that of the clinical-pathologic model (P < .001). In addition, the novel nomogram incorporating these variables showed good discrimination (AUC, 0.90) and calibration abilities (calibration slope, 0.91) in the independent validation cohort.The results of the study by Kim et al are also in agreement with recently published nomograms incorporating MRI features and using texture analysis or machine learning–based radiomics. In recent years, several studies have proposed different techniques for quantifying tumor heterogeneity and the irregularity of tissue components. These techniques are based on statistical and transform-based methods. The most recent studies had excellent performances in predicting pCR (AUC range, 0.78–0.93) (10). In particular, Sutton et al (10) used a semiautomated volumetric tumor segmentation technique that provides more reproducible radiomics features. This improved machine learning classifier accuracy in combination with breast cancer molecular subtype and resulted in a significant improvement in classification performance for detecting pCR compared with molecular subtype alone.However, as Kim et al point out, although their nomogram does not include a machine learning method combining radiomic or texture analysis models with molecular subtypes, texture analysis and radiomics are not easily implementable into clinical practice, and reproducibility may be low. The MRI variables used in their current work, conversely, are relatively simple and easy to apply without the need for dedicated software or sophisticated postprocessing.Another important finding of the study by Kim et al is that the absence of enhancement in the tumor bed, small tumor size, and low lesion-to-background parenchymal signal enhancement ratio at MRI were independently associated with pCR. This suggests that mild enhancement in the tumor bed might represent pCR status due to chemotherapy-induced changes, as previously demonstrated (7). Therefore, the novel nomogram proposed by Kim et al can quantify the probability of pCR based on residual tumor size and lesion-to-background parenchymal signal enhancement ratio at post-NAC MRI by regarding them as continuous variables, rather than as dichotomized parameters of the absence or presence of residual enhancement. Other major strengths of the study are the inclusion of all molecular subtypes, the large sample size, and the use of a multivariable statistical analysis.Some important caveats of this study must be acknowledged—in particular, the retrospective data collection and the use of a single-center university hospital without external validation with data from other institutions. These study characteristics indicate that the nomogram proposed, if further validated with a randomized multicenter study, could be easily integrated into a clinical diagnostic workflow.There is also no mention of the axillary response after NAC, which is crucial for proper axillary management. Several articles have discussed the role of sentinel node biopsy after NAC in patients with biopsy-proven node-positive cancer, representing a valid and powerful alternative approach to complete axillary lymph node dissection. As Kim et al point out in the Discussion, this represents a possible limitation of the study; however, the authors are planning a prospective trial to evaluate the safety of breast surgery omission after NAC in select patients. In the trial, sentinel lymph node biopsy or axillary lymph node dissection will be performed according to individual axillary status.In conclusion, Kim et al propose an innovative and easy-to-apply nomogram incorporating MRI and clinical-pathologic parameters for accurate prediction of pCR after NAC. This nomogram could be used for a better stratification of patients with breast cancer and to determine the most appropriate surgical management after NAC, although further external validation is required. Future studies are needed to verify whether the nomogram proposed by Kim et al could be further improved by integrating radiomics or texture analysis methods.Disclosures of Conflicts of Interest: M.I. disclosed no relevant relationships. A.P. disclosed no relevant relationships.References1. Cortazar P, Zhang L, Untch M, Pathological complete response and long-term clinical benefit in breast cancer: the CTNeoBC pooled analysis. Lancet 2014;384(9938):164–172. Crossref, Medline, Google Scholar2. Teshome M, Hunt KK. Neoadjuvant therapy in the treatment of breast cancer. Surg Oncol Clin N Am 2014;23(3):505–523. Crossref, Medline, Google Scholar3. Kim SY, Cho N, Choi Y, Factors affecting pathologic complete response following neoadjuvant chemotherapy in breast cancer: development and validation of a predictive nomogram. Radiology 2021. https://doi.org/10.1148/radiol.2021203871 Published online March 23, 2021. Google Scholar4. Colleoni M, Bagnardi V, Rotmensz N, A nomogram based on the expression of Ki-67, steroid hormone receptors status and number of chemotherapy courses to predict pathological complete remission after preoperative chemotherapy for breast cancer. Eur J Cancer 2010;46(12):2216–2224. Crossref, Medline, Google Scholar5. Marinovich ML, Macaskill P, Irwig L, Agreement between MRI and pathologic breast tumor size after neoadjuvant chemotherapy, and comparison with alternative tests: individual patient data meta-analysis. BMC Cancer 2015;15(1):662. Crossref, Medline, Google Scholar6. Chen X, Chen X, Yang J, Li Y, Fan W, Yang Z. Combining Dynamic Contrast-Enhanced Magnetic Resonance Imaging and Apparent Diffusion Coefficient Maps for a Radiomics Nomogram to Predict Pathological Complete Response to Neoadjuvant Chemotherapy in Breast Cancer Patients. J Comput Assist Tomogr 2020;44(2):275–283. Crossref, Medline, Google Scholar7. Kim SY, Cho N, Shin SU, Contrast-enhanced MRI after neoadjuvant chemotherapy of breast cancer: lesion-to-background parenchymal signal enhancement ratio for discriminating pathological complete response from minimal residual tumour. Eur Radiol 2018;28(7):2986–2995. Crossref, Medline, Google Scholar8. Kim SY, Cho N, Park IA, Dynamic contrast-enhanced breast MRI for evaluating residual tumor size after neoadjuvant chemotherapy. Radiology 2018;289(2):327–334. Link, Google Scholar9. Keam B, Im SA, Park S, Nomogram predicting clinical outcomes in breast cancer patients treated with neoadjuvant chemotherapy. J Cancer Res Clin Oncol 2011;137(9):1301–1308. Crossref, Medline, Google Scholar10. Sutton EJ, Onishi N, Fehr DA, A machine learning model that classifies breast cancer pathologic complete response on MRI post-neoadjuvant chemotherapy. Breast Cancer Res 2020;22(1):57. Crossref, Medline, Google ScholarArticle HistoryReceived: Jan 17 2021Revision requested: Feb 01 2021Revision received: Feb 03 2021Accepted: Feb 05 2021Published online: Mar 23 2021Published in print: May 2021 FiguresReferencesRelatedDetailsAccompanying This ArticleFactors Affecting Pathologic Complete Response Following Neoadjuvant Chemotherapy in Breast Cancer: Development and Validation of a Predictive NomogramMar 23 2021RadiologyRecommended Articles Preoperative Axillary US in Early-Stage Breast Cancer: Potential to Prevent Unnecessary Axillary Lymph Node DissectionRadiology2018Volume: 288Issue: 1pp. 55-63Use of Pretreatment Breast MRI to Predict Failed Sentinel Lymph Node Identification after Neoadjuvant ChemotherapyRadiology2020Volume: 295Issue: 2pp. 283-284Can We Use MRI and US to Predict Axillary Node Response in Breast Cancer?Radiology2019Volume: 293Issue: 1pp. 58-59Twinkling: A Useful Adjunct for Identifying Biopsy Clips on US ImagesRadiology: Imaging Cancer2023Volume: 5Issue: 4Axillary Nodal Evaluation in Breast Cancer: State of the ArtRadiology2020Volume: 295Issue: 3pp. 500-515See More RSNA Education Exhibits How to Perform CT-Guided Wire and Wireless Localization of Previously Positive Axillary Lymph Nodes (ALN) after Neoadjuvant Chemotherapy (NACT) for Breast Cancer: Pearls and PitfallsDigital Posters2020How to Navigate Breast Tumor Board: A Resident and Fellow PrimerDigital Posters2022Axillary Imaging in Breast Cancer: When, Who and How?Digital Posters2022 RSNA Case Collection Occult Breast CancerRSNA Case Collection2022Inflammatory breast cancerRSNA Case Collection2020Adenoid Cystic Carcinoma RSNA Case Collection2021 Vol. 299, No. 2 Metrics Altmetric Score PDF download