piNET: An Automated Proliferation Index Calculator Framework for Ki67 Breast Cancer Images
Bibliographic record
Abstract
Robust algorithms which are generalizable to multicenter datasets have large potential in digital histopathology. Combatting challenges such as different staining protocols, images, stain and scanner vendors, provide significant value to diagnostic pathology. Automated workflows have the potential to decrease turnaround time, improve efficiency and accuracy. In this work, two different frameworks for proliferation index quantification were developed and analyzed. First, a novel unsupervised color separation pipeline based on the IHC color histogram was proposed for the robust analysis of Ki67 and hematoxylin stained images in multicentre datasets. An “overstaining” threshold was implemented to adjust for background overstaining, and an automated nuclei radius estimator is designed to improve nuclei detection. After implementing this method, we found it to perform poorly on clinical data, specifically the nuclei detection portion. Therefore, this initiated the development of piNET, an automated Proliferation Index Calculator for Ki67 stained digital pathology images. The objective of this model was generalizability, minimal ground truth generation and robustness to multi-center datasets. In this work, a novel pipeline, piNET, can detect immuno-positive and immuno-negative tumor cells, which aids in the quantification of proliferation index. This pipeline is robust to multicenter data, has been assessed on five datasets, in order to accurately claim that the pipeline is generalizable. The F1 Scores proliferation index and classifications were thoroughly evaluated on multicenter data, 173 patched data, 90 tissue micro-arrays and 55 whole slide images. This pipeline was built on the U-NET, in combination with regression-based modeling and using a novel data partition approach. The piNET, trained on a single dataset, obtained an accuracy rate of 86% on the tissue micro-array, and 89% on patched data, and 76% on whole slide images. The proposed method can achieve an overall accuracy rate of 85% across five datasets and a Proliferation Index difference and R2 of 5.6% and 0.84 across four datasets respectively.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.005 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.006 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".