A secondary analysis of the phase III APHINITY trial evaluated manual, digital, and AI-driven tumour-infiltrating lymphocyte scoring methods, assessing their comparative predictive utility in HER2-positive early breast cancer. Findings suggest AI-based TIL quantification may offer clinically meaningful prognostic and predictive information in this setting.
A secondary analysis published in The Lancet Oncology examined tumour-infiltrating lymphocyte (TIL) scoring across three methodologies—manual pathologist assessment, digital image analysis, and artificial intelligence-based quantification—using biospecimens from the phase III APHINITY randomised trial. The analysis focused on the predictive value of each approach in the context of adjuvant pertuzumab plus trastuzumab-based chemotherapy for HER2-positive early breast cancer.
TILs have emerged as a biologically relevant immune biomarker in breast oncology, with higher stromal TIL density associated with improved outcomes in HER2-positive and triple-negative subtypes. The APHINITY trial previously established the benefit of adding pertuzumab to trastuzumab-based adjuvant regimens; this secondary analysis extends that work by interrogating whether TIL scoring methodology influences the biomarker's predictive performance for treatment benefit or survival endpoints.
The inclusion of AI-driven scoring alongside conventional manual and digital approaches is clinically significant, as it raises the possibility of standardising TIL assessment at scale—addressing longstanding concerns about interobserver variability in manual scoring. If AI-derived TIL scores demonstrate concordant or superior predictive accuracy, they could support more consistent biomarker-driven decision-making in HER2-positive adjuvant therapy planning.
Full quantitative results, including concordance statistics and outcome correlations across scoring methods, were not available in the indexed source data at the time of publication. Clinicians should consult the complete manuscript in The Lancet Oncology for endpoint-specific findings and methodological details.
Key points
- Secondary analysis of the phase III APHINITY trial compared manual, digital, and AI-based TIL scoring in HER2-positive early breast cancer.
- The analysis assessed the predictive value of each TIL quantification method in the setting of adjuvant pertuzumab plus trastuzumab-based chemotherapy.
- AI-driven TIL scoring could address interobserver variability inherent in manual pathological assessment, potentially enabling more standardised biomarker evaluation.
- TIL density is an established immune biomarker associated with prognosis and treatment response in HER2-positive breast cancer.
- Complete quantitative outcome data were not available in the indexed abstract; full results are published in The Lancet Oncology.
Why it matters
Standardising TIL scoring through validated AI-based methods could improve the reproducibility of immune biomarker assessment and refine patient selection for intensified adjuvant anti-HER2 regimens. This has direct implications for oncologists managing HER2-positive early breast cancer in routine clinical practice.