Early prediction of neoadjuvant chemotherapy response for advanced breast cancer using PET/MRI image deep learning

This study aimed to investigate the predictive efficacy of positron emission tomography/computed tomography (PET/CT) and magnetic resonance imaging (MRI) for the pathological response of advanced breast cancer to neoadjuvant chemotherapy (NAC). The breast PET/MRI image deep learning model was introduced and compared with the conventional methods. PET/CT and MRI parameters were evaluated before and after the first NAC cycle in patients with advanced breast cancer [n = 56; all women; median age, 49 (range 26–66) years]. The maximum standardized uptake value (SUVmax), metabolic tumor volume (MTV), and total lesion glycolysis (TLG) were obtained with the corresponding baseline values (SUV0, MTV0, and TLG0, respectively) and interim PET images (SUV1, MTV1, and TLG1, respectively). Mean apparent diffusion coefficients were obtained from baseline and interim diffusion MR images (ADC0 and ADC1, respectively). The differences between the baseline and interim parameters were measured (ΔSUV, ΔMTV, ΔTLG, and ΔADC). Subgroup analysis was performed for the HER2-negative and triple-negative groups. Datasets for convolutional neural network (CNN), assigned as training (80%) and test datasets (20%), were cropped from the baseline (PET0, MRI0) and interim (PET1, MRI1) images. Histopathologic responses were assessed using the Miller and Payne system, after three cycles of chemotherapy. Receiver operating characteristic curve analysis was used to assess the performance of the differentiating responders and non-responders. There were six responders (11%) and 50 non-responders (89%). The area under the curve (AUC) was the highest for ΔSUV at 0.805 (95% CI 0.677–0.899). The AUC was the highest for ΔSUV at 0.879 (95% CI 0.722–0.965) for the HER2-negative subtype. AUC improved following CNN application (SUV0:PET0 = 0.652:0.886, SUV1:PET1 = 0.687:0.980, and ADC1:MRI1 = 0.537:0.701), except for ADC0 (ADC0:MRI0 = 0.703:0.602). PET/MRI image deep learning model can predict pathological responses to NAC in patients with advanced breast cancer.

[1]  Z. Hall.  Cancer , 1906, The Hospital.

[2]  D. Hubel,et al.  Receptive fields, binocular interaction and functional architecture in the cat's visual cortex , 1962, The Journal of physiology.

[3]  B. Meyer,et al.  [Results and indications of cochlear implant in 19 cases of total pre-speech deafness]. , 1990, Annales d'oto-laryngologie et de chirurgie cervico faciale : bulletin de la Societe d'oto-laryngologie des hopitaux de Paris.

[4]  Z. Baloch,et al.  Pathologic response to induction chemotherapy in locally advanced carcinoma of the breast: a determinant of outcome. , 1995, Journal of the American College of Surgeons.

[5]  A. Hutcheon,et al.  A new histological grading system to assess response of breast cancers to primary chemotherapy: prognostic significance and survival. , 2003, Breast.

[6]  C. Sotak.  Nuclear magnetic resonance (NMR) measurement of the apparent diffusion coefficient (ADC) of tissue water and its relationship to cell volume changes in pathological states , 2004, Neurochemistry International.

[7]  R. Bast,et al.  American Society of Clinical Oncology 2007 update of recommendations for the use of tumor markers in breast cancer. , 2007, Journal of clinical oncology : official journal of the American Society of Clinical Oncology.

[8]  D. Collins,et al.  Diffusion-weighted MRI in the body: applications and challenges in oncology. , 2007, AJR. American journal of roentgenology.

[9]  S. Paik,et al.  Preoperative chemotherapy: updates of National Surgical Adjuvant Breast and Bowel Project Protocols B-18 and B-27. , 2008, Journal of clinical oncology : official journal of the American Society of Clinical Oncology.

[10]  J. Gralow,et al.  Neoadjuvant chemotherapy for locally advanced breast cancer. , 2009, Seminars in radiation oncology.

[11]  E. Merkle,et al.  Short- and midterm reproducibility of apparent diffusion coefficient measurements at 3.0-T diffusion-weighted imaging of the abdomen. , 2009, Radiology.

[12]  S. Gautam,et al.  Identification of residual breast carcinoma following neoadjuvant chemotherapy: diffusion-weighted imaging--comparison with contrast-enhanced MR imaging and pathologic findings. , 2010, Radiology.

[13]  Giovanni Seni,et al.  Ensemble Methods in Data Mining: Improving Accuracy Through Combining Predictions , 2010, Ensemble Methods in Data Mining.

[14]  Qifeng Yang,et al.  Meta-analysis confirms achieving pathological complete response after neoadjuvant chemotherapy predicts favourable prognosis for breast cancer patients. , 2011, European journal of cancer.

[15]  M. Hatt,et al.  Comparison Between 18F-FDG PET Image–Derived Indices for Early Prediction of Response to Neoadjuvant Chemotherapy in Breast Cancer , 2013, The Journal of Nuclear Medicine.

[16]  S. Rodenhuis,et al.  FDG PET/CT during neoadjuvant chemotherapy may predict response in ER-positive/HER2-negative and triple negative, but not in HER2-positive breast cancer. , 2013, Breast.

[17]  M. Miyazaki,et al.  Diffusion-weighted imaging reflects pathological therapeutic response and relapse in breast cancer , 2014, Breast Cancer.

[18]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[19]  S. Rodenhuis,et al.  Combined use of 18F-FDG PET/CT and MRI for response monitoring of breast cancer during neoadjuvant chemotherapy , 2014, European Journal of Nuclear Medicine and Molecular Imaging.

[20]  Guangyu Liu,et al.  18F-fluorodeoxyglucose (FDG) PET/CT after two cycles of neoadjuvant therapy may predict response in HER2-negative, but not in HER2-positive breast cancer , 2015, Oncotarget.

[21]  M. Hatt,et al.  Early Metabolic Response to Neoadjuvant Treatment: FDG PET/CT Criteria according to Breast Cancer Subtype. , 2015, Radiology.

[22]  V. Goh,et al.  Predicting Response to Neoadjuvant Chemotherapy with PET Imaging Using Convolutional Neural Networks , 2015, PloS one.

[23]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[24]  Sungeun Kim,et al.  Early prediction of pathological complete response in luminal B type neoadjuvant chemotherapy-treated breast cancer patients: comparison between interim 18F-FDG PET/CT and MRI , 2015, Nuclear medicine communications.

[25]  Jae Y. Shin,et al.  Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning? , 2016, IEEE transactions on medical imaging.

[26]  C. Park,et al.  Pathologic Evaluation of Breast Cancer after Neoadjuvant Therapy , 2016, Journal of pathology and translational medicine.

[27]  Nima Tajbakhsh,et al.  Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning? , 2016, IEEE Transactions on Medical Imaging.

[28]  S. Sheikhbahaei,et al.  FDG-PET/CT and MRI for Evaluation of Pathologic Response to Neoadjuvant Chemotherapy in Patients With Breast Cancer: A Meta-Analysis of Diagnostic Accuracy Studies. , 2016, The oncologist.

[29]  June-Goo Lee,et al.  Deep Learning in Medical Imaging: General Overview , 2017, Korean journal of radiology.

[30]  Nico Karssemeijer,et al.  Large scale deep learning for computer aided detection of mammographic lesions , 2017, Medical Image Anal..

[31]  B. Erickson,et al.  Machine Learning for Medical Imaging. , 2017, Radiographics : a review publication of the Radiological Society of North America, Inc.

[32]  Lubomir M. Hadjiiski,et al.  Bladder Cancer Treatment Response Assessment in CT using Radiomics with Deep-Learning , 2017, Scientific Reports.

[33]  Qianni Zhang,et al.  Three-Class Mammogram Classification Based on Descriptive CNN Features , 2017, BioMed research international.

[34]  Wen Gao,et al.  Diffusion-weighted imaging in monitoring the pathological response to neoadjuvant chemotherapy in patients with breast cancer: a meta-analysis , 2018, World Journal of Surgical Oncology.

[35]  Atsuto Maki,et al.  A systematic study of the class imbalance problem in convolutional neural networks , 2017, Neural Networks.

[36]  Ulas Bagci,et al.  Deep learning beyond cats and dogs: recent advances in diagnosing breast cancer with deep neural networks. , 2018, The British journal of radiology.

[37]  Aly A. Valliani,et al.  Deep Learning and Neurology: A Systematic Review , 2019, Neurology and Therapy.

[38]  Mohammad Sohel Rahman,et al.  MultiResUNet : Rethinking the U-Net Architecture for Multimodal Biomedical Image Segmentation , 2019, Neural Networks.

[39]  Jeremiah Neubert,et al.  Deep learning approaches to biomedical image segmentation , 2020 .