Prediction of pathologic complete response to neoadjuvant systemic therapy in triple negative breast cancer using deep learning on multiparametric MRI

Abstract

Triple-negative breast cancer (TNBC) is an aggressive subtype of breast cancer. Neoadjuvant systemic therapy (NAST) followed by surgery are currently standard of care for TNBC with 50-60% of patients achieving pathologic complete response (pCR). We investigated ability of deep learning (DL) on dynamic contrast enhanced (DCE) MRI and diffusion weighted imaging acquired early during NAST to predict TNBC patients’ pCR status in the breast. During the development phase using the images of 130 TNBC patients, the DL model achieved areas under the receiver operating characteristic curves (AUCs) of 0.97 ± 0.04 and 0.82 ± 0.10 for the training and the validation, respectively. The model achieved an AUC of 0.86 ± 0.03 when evaluated in the independent testing group of 32 patients. In an additional prospective blinded testing group of 48 patients, the model achieved an AUC of 0.83 ± 0.02. These results demonstrated that DL based on multiparametric MRI can potentially differentiate TNBC patients with pCR or non-pCR in the breast early during NAST.

Artificial Intelligence-Enhanced Breast MRI

“…Subsequently, in 2020, El Adoui et al 87 created a CNN based on multi-input DCE-MRI data before and after the first cycle of NAC in 42 patients with breast cancer, achieving an AUC of 0.91 for predicting pCR. In a similar fashion, Zhou et al 93 in 2023 explored the use of multichannel 3D CNNs (instead of branches) for predicting pCR. However, instead of using DCE-MRI data only as in the studies by El Adoui et al, Zhou et al incorporated multiparametric MRI (DCE-MRI and DWI) data before and after 4 cycles of NAC.…”

Multiparametric Approach to Breast Cancer With Emphasis on Magnetic Resonance Imaging in the Era of Personalized Breast Cancer Treatment

“…However, multiparametric data are now also used to tackle more complex tasks such as predicting subtypes and treatment response after neoadjuvant treatment. We have summarized selected recent research investigating the prediction of subtypes 10,27,29,30,91,[103][104][105][106][107][108][109] (Table 2), response after neoadjuvant treatment [110][111][112] (Table 3), and evaluating TILs or CAF [113][114][115][116] (Table 4) to provide an overview of the trends in parameters and methods used for analysis. The most popular combinations of imaging include DCE, DWI (ADC value), and T2WI.…”

“…Intraindividual changes between images of different time points (delta) can be used as additional parameters. 111 Excellent prediction performances with AUC over 0.9 are achieved, [110][111][112] although they became lower when tested on prospective external test set. 112 Multiparametric approach may help in evaluating these status reflecting TME.…”

“…111 Excellent prediction performances with AUC over 0.9 are achieved, [110][111][112] although they became lower when tested on prospective external test set. 112 Multiparametric approach may help in evaluating these status reflecting TME. [113][114][115][116]…”

Deep Learning for Fully Automatic Tumor Segmentation on Serially Acquired Dynamic Contrast-Enhanced MRI Images of Triple-Negative Breast Cancer

“…In addition, we calculated and used the following three semiquantitative maps: positive enhancement integral (PEI) [40], signal enhancement ratio (SER) [41], and maximum slope of increase (MSI). These maps [42] were calculated on an AW Server using the software provided by the vendor (v3.2, GE Healthcare, Milwaukee, WI, USA).…”