Exposure to US Cancer Drugs With Lack of Confirmed Benefit After US Food and Drug Administration Accelerated Approval

This cross-sectional study evaluates patient exposure to oncology drugs withdrawn from the US Food and Drug Administration (FDA) Accelerated Approval program.

We estimate the proportion of patients who received treatment for 5 oncologyrelated indications later withdrawn after failure to confirm efficacy under the AA program.
Methods | This cross-sectional study included patients with advanced or recurrent breast, bladder, hepatocellular, gastric, or small cell lung cancer between May 18, 2016, andMarch 8, 2022, treated with at least 1 line of systemic therapy. We used deidentified, electronic health record (EHR)-derived, patientlevel data curated via technology-enabled abstraction. 4 The Copernicus Group and University of Pennsylvania institutional review boards approved this study and granted waivers of informed consent owing to use of deidentified data. This study followed the STROBE reporting guideline.
We studied 5 disease-specific AA indications with subsequent published negative phase III confirmatory trials and indication withdrawal (Table). These drugs had additional biomarkerspecific, cancer-agnostic indication approvals, such that they were available for use before AA and continued to be available after AA withdrawal. Outcomes were calculated for patients aged 18 years or older who initiated therapy and were eligible for the AA indication. Race and ethnicity were identified from the EHR and included as covariates because minority status is associated with decreased access to novel therapies. We excluded patients with 1 or more cancers or, for biomarker-specific indications, missing biomarker data. The primary outcome was initiation of AA therapies as a proportion of all indication-specific treatment initiations. We calculated the primary outcome at 3 intervals: AA to indication withdrawal, AA to negative confirmatory trial, and negative trial to indication withdrawal. Because 4 indications were reviewed at the April 2021 FDA Oncology Drug Advisory Committee meeting, we included this date as a discrete point. Logistic regression-estimated time trends with restricted cubic splines were used to estimate the daily percentage of treatmenteligible patients who initiated a withdrawn AA drug. Analyses were performed using SAS, version 9.4 (SAS Institute Inc).  Figure). Prevalence of AA drug initiations was higher between AA and negative trial publication (overall, 35.5%; triple-negative breast, 24.1%; bladder, 39.7%; hepatocellular, 55.6%; gastric, 71.4%; small cell lung, 21.0%) than between negative trial publication and withdrawal (15.7%, 8.6%, 12.8%, 20.3%, 38.6%, 2.5%, respectively) (Table).

Results
Discussion | Among 5 oncology indications, 26.1% of eligible treatment initiations involved an AA indication subsequently withdrawn due to lack of benefit. An expected trade-off exists between expediting access to promising cancer drugs and withdrawal of some indications. 5 Given the growth of withdrawals due to negative confirmatory trials and emerging evidence on the high spending associated with AA drugs, it is critical to balance early access against population-level exposure to cancer therapies with no benefit over standard of care. 6 Limitations included an inability to assess population-level exposure because only 5 withdrawn AA indications had sufficient sample and follow-up for analysis. Earlier access and more rapid FDA responses to negative confirmatory trial data, a key proposal of the Accelerated Approval Integrity Act proposed in March 2022, may minimize exposure to AA therapies with lack of benefit.   Observed quarterly prescription rate Estimated prescription rate Shown are the trends in use of oncology drugs that received AA and were subsequently withdrawn between 2016 and 2022. Observed quarterly prescription rate of AA therapies is a proportion of all treatment initiations for an indication, and estimated prescription rate is the daily percentage of treatment-eligible patients who used an AA drug based on logistic regression with restricted cubic splines. ODAC indicates the US Food and Drug Administration Oncology Drug Advisory Committee.

Increased Tumor Mutation Burden Levels and Sensitivity of Non-Small Cell Lung Cancer to PD-L1 Blockade
To the Editor In their cohort study in JAMA Oncology, Ricciuti et al 1 found that increased tumor mutation burden levels were associated with immune cell infiltration and inflammatory T-cell-mediated response, resulting in increased sensitivity to programmed cell death-1 (PD-1) or PD-1 ligand (PD-L1) blockade in non-small cell lung cancer (NSCLC) across PD-L1 expression subgroups. These findings provide new insights into the use of PD-1/PD-L1 inhibitors in advanced NSCLC. However, we have several concerns. First, according to the data of a phase 2 clinical trial, 2 PD-1 blockade monotherapy was used for first-line treatment of EGFRmutant NSCLC, and the proportion of patients with PD-L1 expression 50% or greater accounted for 73%, while the objective response rate was only 9%. The clinical trial was terminated early due to lack of efficacy. In the retrospective study by Ricciuti et al, 1 138 of 143 (96.5%) patients with NSCLC with EGFR mutation were defined as having low tumor mutation burden, and among them, patients with PD-L1 expression 50% or greater had an objective response rate as high as 38.1% (95% CI, 33.3%-43.0%). Therefore, we hope Ricciuti et al provide survival data of patients with common driver alterations independently to prove the authenticity of the data.
Second, more than 65% of the enrolled patients in the present study 1 had previously received 1 or more antitumor therapies, and it was inevitable that this would affect overall survival (OS). Multiple clinical trials have demonstrated that first-line treatment with tyrosine kinase inhibitors significantly prolonged OS in patients with positive driver alterations. 3,4 Therefore, we consider that the published results of this study 1 cannot exclude OS bias caused by inclusion screening. We recommend that Ricciuti et al provide survival data and duration of remission after initiation of immunotherapy for patients who did not receive immune checkpoint inhibitors in the first-line therapy, to provide meaningful guidance for clinical treatment selection.
In addition, we found a small mistake in panel C of Figure