Technical validation of PD-L1 SP142 assay for TNBC in a large academic center: Unexpected challenges.
Bibliographic record
Abstract
e15253 Background: The IMpassion 130 trial reported that 41-46% of TNBC are positive for PD-L1 SP142 based on centralized testing. While the FDA approved it as a companion diagnostic assay with cut off value of 1%, there are no published guidelines for technical validation in local laboratories. Furthermore, with the need to evaluate staining only in inflammatory cells (IC) the readout becomes particularly challenging for cases around the cutoff point. We report our prevalence of PD-L1 positivity and proportion of borderline cases in our validation. Methods: We ran PDL1 SP142 using the recommended protocol from Roche on the following groups: G1 - 53 consecutive TNBC cases (excisions tumors ≥7mm), G2 - 10 consecutive TNBC cases (breast specimens only) that would have met the inclusion criteria for the trial (LABC with poor NAT response or metastatic), G3 – proficiency testing slide from an EQA provider (4 cores of TNBC 1 negative, 1 borderline low positive and 2 high positive). All cases were read by 2 breast pathologists that received training on PDL1 readout by Roche (one – in class and online, one – online). Results: Borderline staining around the 1% cutoff (11 of 63) and low positive (9 of 63) were encountered in the validation sets more often than shown in training sessions. Classification of borderline cases may have a substantial impact on the rate of PDL1 positivity that potentially could approach 75%. Borderline cases often have IC staining either at the advancing tumor edge, as small collections of highly positive IC or around benign ducts entrapped within the tumor. Conclusions: The prevalence of PDL1 positive TNBC in all comers could be higher than observed in IMpassion 130 (at least 50-60%). Pathologist training in readout is of utmost importance due to high rate of borderline/low positive cases raising the need for central testing in borderline cases. Guideline recommendations for technical validation similar to other class II biomarkers would be useful. Studies on interobserver variability are needed to further address this potential issue. [Table: see text]
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 | 0.069 | 0.054 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.003 | 0.001 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.003 | 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".