Improving Patient Outcomes

We thank David Rimm for his comments and the opportunity to respond. As a cancer community unified by a common mission, we have a great responsibility to tackle serious unmet patient needs and deliver effective treatments through innovative and collaborative approaches to drug and diagnostic discovery, development, and integration into clinical practice.

Patients with advanced or metastatic triple-negative breast cancer (TNBC) experience poor outcomes relative to patients with other breast cancer subtypes (1,2). TECENTRIQ (atezolizumab) in combination with nab-paclitaxel was granted FDA Accelerated Approval for the treatment of adult patients with unresectable, locally advanced, or metastatic TNBC whose tumors express PD-L1 – defined as PD-L1 stained tumor-infiltrating immune cells (IC) of any intensity covering ≥ 1 percent of the tumor area – as determined by an FDA-approved test (3). The VENTANA PD-L1 (SP142) Assay (SP142 assay) is the FDA-approved companion diagnostic to TECENTRIQ (4)*. These approvals were based on the IMpassion130 clinical trial (5) and represent the first immunotherapy regimen for breast cancer and an important new treatment for its most aggressive subtype.

In the IMpassion130 study, patients whose tumors were positive for PD-L1 showed a stratified hazard ratio (HR) of 0.62 (p<0.0001) for median progression free survival (PFS) in favor of the TECENTRIQ + nab-paclitaxel combination. Exploratory analysis showed no benefit in PD-L1-negative patients as tested by the SP142 assay, further supporting the assay’s role in identifying patients who may benefit from the TECENTRIQ combination. The second interim overall survival (OS) analysis demonstrated a HR of 0.71 in PD-L1-positive patients; OS was not formally tested in the PD-L1-positive population due to the hierarchical design for OS. As with PFS, exploratory analyses showed no benefit in the PD-L1-negative patients (6). Additional exploratory biomarker analysis evaluating PD-L1 expression on tumor cells, stromal tumor-infiltrating immune cells, and cytotoxic T cells concluded that PD-L1 IC expression based on the SP142 assay was the best predictor of clinical benefit (7).

Regarding the differential performance characteristics of PD-L1 assays, it is important to distinguish analytic sensitivity and specificity from the ability of these assays to predict treatment response. The SP142 assay incorporates an amplified detection system with a different staining pattern to other approved PD-L1 assays. This has created challenges for pathologists seeking to harmonize these assays so that they can use one PD-L1 assay to make therapeutic decisions for several different PD-1 and PD-L1 inhibitors across several different indications based on different scoring algorithms. Although the desire for streamlining of testing is understood, the role of a companion diagnostic assay is to discriminate between responders and non-responders for that specific therapeutic product in a specific indication, with a cutoff based on clinical outcomes.

In vitro diagnostic devices (IVDs) are subject to design controls and must be validated and comply with Quality Systems regulations. Companion diagnostic assays must demonstrate both analytic and clinical validity. Results of validation studies approved by the FDA pertaining to the analytical specificity and sensitivity, repeatability, precision, and readability of the SP142 assay are publicly available (8).

The era of precision medicine in oncology reinforces the role of pathologists as partners to oncologists, and Roche wholeheartedly supports the education and training of the global pathology community. Data from the Roche international pathologist training program for the SP142 assay (8) have demonstrated reproducibility in both NSCLC (TC/IC algorithm) and urothelial carcinoma (UC) (IC only algorithm) indications. Average agreement rates were 88.3 and 95.3 percent respectively (9), likely reflecting the simpler IC-only algorithm across two categories in UC. The TNBC algorithm similarly includes IC-only assessment across two categories. Roche has currently trained more than 1,000 pathologists globally through in-person programs for non-small cell lung cancer (NSCLC), UC and TNBC indications; this data will be published shortly.

Although NSCLC is not a surrogate for TNBC, the Blueprint 2 study (10) conducted in NSCLC noted that the SP142 assay was the only assay to show moderate-to-strong agreement between pathologists versus the trainer for distinguishing IC0 versus IC1, 2, and 3, which corresponds with the IC1 percent cutoff.

Roche is dedicated to working with pathologists and the healthcare community to find solutions to challenges in an evolving era of precision medicine and remains committed to advancing scientific innovation to improve clinical outcomes.

*The VENTANA PD-L1 (SP142) Assay may not be available for the TNBC indication in all geographies.