CoChat
Sign Up
Cited by 113AI in cancer detectionBreast Cancer Treatment StudiesRadiomics and Machine Learning in Medical Imaging

The DOI matches its Crossref record; 48 of the 49 references it cites were found in a registry or an index.

Deep learning-based predictive biomarker of pathological complete response to neoadjuvant chemotherapy from histological images in breast cancer

Open access
Li, Fengling, Yang, Yongquan, Wei, Yani, He, Ping, Chen, Jie, Zheng, Zhongxi, Bu, Hong·2021·Journal of Translational Medicine
DOI 10.1186/s12967-021-03020-zPMC PMC8365907
Open this paper in CoChat — free accountOpen PDF

Abstract

Background: Pathological complete response (pCR) is considered a surrogate endpoint for favorable survival in breast cancer patients treated with neoadjuvant chemotherapy (NAC). Predictive biomarkers of treatment response are crucial for guiding treatment decisions. With the hypothesis that histological information on tumor biopsy images could predict NAC response in breast cancer, we proposed a novel deep learning (DL)-based biomarker that predicts pCR from images of hematoxylin and eosin (H&E)-stained tissue and evaluated its predictive performance. Methods: In total, 540 breast cancer patients receiving standard NAC were enrolled. Based on H&E-stained images, DL methods were employed to automatically identify tumor epithelium and predict pCR by scoring the identified tumor epithelium to produce a histopathological biomarker, the pCR-score. The predictive performance of the pCR-score was assessed and compared with that of conventional biomarkers including stromal tumor-infiltrating lymphocytes (sTILs) and subtype. Results: The pCR-score derived from H&E staining achieved an area under the curve (AUC) of 0.847 in predicting pCR directly, and achieved accuracy, F1 score, and AUC of 0.853, 0.503, and 0.822 processed by the logistic regression method, respectively, higher than either sTILs or subtype; a prediction model of pCR constructed by integrating sTILs, subtype and pCR-score yielded a mean AUC of 0.890, outperforming the baseline sTIL-subtype model by 0.051 (0.839, P = 0.001). Conclusion: The DL-based pCR-score from histological images is predictive of pCR better than sTILs and subtype, and holds the great potentials for a more accurate stratification of patients for NAC.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s12967-021-03020-z

Cites

48 of 48 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig. 1 The pipeline of the pCR-score computing consists of five sub-steps: a pre-processing; b CNN I; c middle-processing; d CNN II; and e post-processing. First, the pre-processing step segments valid tissue areas from the input WSI and crops the segmented valid tissue areas into small tiles. Second, the CNN I takes the cropped tiles as inputs and identifies TE regions by mapping the input tiles into probabilities corresponding to TE. Third, the middle-processing step selects TE tiles with identified high probabilities from the outputs of CNN I. Fourth, the CNN II takes the selected TE tiles as inputs and score the pCR of the input tiles by mapping them into probabilities corresponding to pCR. Finally, the post-processing step fuses the pCR probabilities of TE tiles scored via CNN II to produce the final predicted pCR-score of the input WSI
Fig. 1 The pipeline of the pCR-score computing consists of five sub-steps: a pre-processing; b CNN I; c middle-processing; d CNN II; and e post-processing. First, the pre-processing step segments valid tissue areas from the input WSI and crops the segmented valid tissue areas into small tiles. Second, the CNN I takes the cropped tiles as inputs and identifies TE regions by mapping the input tiles into probabilities corresponding to TE. Third, the middle-processing step selects TE tiles with identified high probabilities from the outputs of CNN I. Fourth, the CNN II takes the selected TE tiles as inputs and score the pCR of the input tiles by mapping them into probabilities corresponding to pCR. Finally, the post-processing step fuses the pCR probabilities of TE tiles scored via CNN II to produce the final predicted pCR-score of the input WSICC BY · PubMed Central
Fig. 2 Examples for the pCR prediction of based on CNNs. A An example of pCR: a shows a represnetative WSI from a patient who achieved the pCR. bThe probability map of TE produced by the CNN I. c The selection of tiles with a high probability of TE. d The map produced by CNN II showing the pCR probability for image a, where pink and white separately represent high and low probabilities of pCR. e The distributions of tile-level pCR scores of the image a. B An example of non-pCR: f shows a representative WSI from a patient who did not achieve the pCR. g, h, and i for image f are corresponding to the steps of b, c, and d for image a. j A predominant distribution of low pCR scores for image f
Fig. 2 Examples for the pCR prediction of based on CNNs. A An example of pCR: a shows a represnetative WSI from a patient who achieved the pCR. bThe probability map of TE produced by the CNN I. c The selection of tiles with a high probability of TE. d The map produced by CNN II showing the pCR probability for image a, where pink and white separately represent high and low probabilities of pCR. e The distributions of tile-level pCR scores of the image a. B An example of non-pCR: f shows a representative WSI from a patient who did not achieve the pCR. g, h, and i for image f are corresponding to the steps of b, c, and d for image a. j A predominant distribution of low pCR scores for image fCC BY · PubMed Central
Fig. 3 ROC curves of raw pCR-scores based on CNN II for pCR prediction in the primary dataset (A) and validation dataset (B)
Fig. 3 ROC curves of raw pCR-scores based on CNN II for pCR prediction in the primary dataset (A) and validation dataset (B)CC BY · PubMed Central
Fig. 4 Comparisons of the pCR prediction performance metrics of sTILs, subtype, and pCR-score in the 16-time repeated validation (In each repeat, we randomly select 60% of the data as training data, and the remaining 40% as testing data. Mean values of each metric were calculated from the 16 repeats to avoid the impact of data bias). A Comparisons of the F1 score and accuracy of models. B Comparisons of the AUCs of models. C Comparisons of the sensitivity (equal to recall score), PPV (equal to precision score), specificity, and NPV of models. D Comparisons of TP, FN, FP, and TN in confusion matrices among models
Fig. 4 Comparisons of the pCR prediction performance metrics of sTILs, subtype, and pCR-score in the 16-time repeated validation (In each repeat, we randomly select 60% of the data as training data, and the remaining 40% as testing data. Mean values of each metric were calculated from the 16 repeats to avoid the impact of data bias). A Comparisons of the F1 score and accuracy of models. B Comparisons of the AUCs of models. C Comparisons of the sensitivity (equal to recall score), PPV (equal to precision score), specificity, and NPV of models. D Comparisons of TP, FN, FP, and TN in confusion matrices among modelsCC BY · PubMed Central
Fig. 5 The distributions of pCR-scores across different subtypes and sTILs densities
Fig. 5 The distributions of pCR-scores across different subtypes and sTILs densitiesCC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 49 references

Cited by

Sorted by citations

Related

See all

Metadata

Authors
Li, F., Yang, Y., Wei, Y., He, P., Chen, J., Zheng, Z., & Bu, H.
Affiliations
Sichuan University (CN)
Corresponding author
  • Zheng, Zhongxi
  • Bu, Hong
ORCID
Venue
Journal of Translational Medicine
Publisher
Springer Science and Business Media LLC
Published
2021-08-16
Volume
19
Issue
1
Pages
348-348
ISSN
1479-5876 (ISSN-L)
DOI
10.1186/s12967-021-03020-z
PMC
PMC8365907
OpenAlex
W3211240094
Cited by
113
References
49
Citations by year
2026: 20 · 2025: 27 · 2024: 34 · 2023: 20 · 2022: 11 · 2021: 1
FWCI
5.83
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Translational Medicine (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • 1·3·5 project for disciplines of excellence (ZYGD18012)
  • the Technological Innovation Project of Chengdu New Industrial Technology Research Institute (2017-CY02–00026-GX)
  • the Sichuan Science and Technology Program (2020YFS0088)
  • the 1·3·5 project for disciplines of excellence Clinical Research Incubation Project (2019HXFH036)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › AI in cancer detection
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Cancer Research › Breast Cancer Treatment Studies
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Radiomics and Machine Learning in Medical Imaging
Keywords
breast cancer, neoadjuvant chemotherapy, pathological complete response, predictive biomarkers, deep learning, stromal tumor-infiltrating lymphocytes, tumor subtypes, treatment response prediction, patient stratification
MeSH terms
Deep Learning, Antineoplastic Combined Chemotherapy Protocols, Breast Neoplasms, Female, Humans, Biomarkers, Lymphocytes, Tumor-Infiltrating, Treatment Outcome, Neoadjuvant Therapy

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review