Abstract 3381: Standardizing the analysis of Ki-67 immunohistochemical assays
Bibliographic record
Abstract
Abstract Intro Immunohistochemical assays of the proliferation marker, Ki-67, have been associated with poorer clinical prognosis in breast cancer. However, a high degree of inconsistency in scores has been demonstrated in inter-laboratory and intra-laboratory Ki-67 positivity scorings, this has limited its potential in clinical practice. In this study, we aim to find a more consistent method for scoring Ki-67 positivity among malignant breast tumours. Methods Six Tissue Microarray (TMA) blocks were sectioned and immunohistochemistry was performed with Anti-Ki-67 antibody. Slides were then evaluated and Ki-67 positive cells in invasive breast carcinoma were scored as a percent positivity manually by a trained analyst with random sample quality assurance (QA) by trained pathologists. This was used as the standard benchmark for the experiment as it has been correlated successfully with clinical outcome. Successively, the same six slides were then annotated on Aperio ePathology software by two observers with different levels of pathology training and experience. The annotated regions were analyzed for Ki-67 positivity with Aperio ePathology software on UHN BioBank servers. The computer analyzed scores were compared to the manual benchmark scores. Results The difference between computer-analyzed and manual-scores were relatively large, Observer-A-annotated-computer-analyzed vs. analyst-manual-scores had a difference of 4.23% to 16.96%, while Observer-B-annotated-computer-analyzed vs. analyst-manual-scores had a difference of 7.13% to 15.03%. Interestingly, Observer-A-annotated-computer-analyzed vs. Observer-B-annotated-computer-analyzed scores only had a difference of 0.49% to 2.91%. Pearson Correlation was calculated for all samples on a case-by-case basis and we found there to be a linear correlation of 0.564, with a P-value of 3.6082×10-8, between the computer scores and the manual scores; suggesting significant correlation between the computer scores and the manual score. Conclusion A significant linear correlation has been observed between the computer score and the manual score. However, while the data does not seem to support the idea that a semi-automatic method of computer scoring will replace analyst manual scoring, most of the large contributing variables have been identified. We plan in the next steps of the project to continue to decrease the effects of such variables. It is interesting that the inter-observer computer score displayed a minimal amount of difference, again with the variables identified. This could signify a more consistent method of Ki-67 scoring. Further experiments will be conducted to continue to reduce the variables and optimize the system to gain similar performance as manual scoring. Hopefully in the near future, computerized immunohistochemical analysis can replace the tedious task of manual scoring. Citation Format: Tian Yu Liu, Trillium Chang, Adewunmi Adeoye, Willa Shi, Sheng-Ben Liang, Dianne Chadwick, Michael H.A. Roehrl, Naomi Miller, Fei-Fei Liu, Susan J. Done. Standardizing the analysis of Ki-67 immunohistochemical assays. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3381. doi:10.1158/1538-7445.AM2015-3381
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 | 0.039 | 0.059 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".