312P - ClinMatch: A clinical trial matching platform that improves trial accessibility among NSCLC patients through comprehensive genomic and clinical profiling

Background

Comprehensive genomic profiling (CGP) has become the cornerstone of personalised oncology treatment. Therapeutic approaches based distinctively to patients’ demographics, prior treatment and genomic profiles are now being investigated in clinical trials. Access to suitable trials have also shown to improve outcome. Current trial subject recruitment process is hampered by manual examination of eligibility criteria. An automated decision support system coupled with CGP may help increase clinical trials recruitment and reduce screen failure rates.

Methods

We developed the ClinMatch system that comprises three cores: (1) A natural language processing algorithm that curates trial descriptions and patient records (2) A cloud analytics platform that processes CGP data (3) a recommendation system that prioritise trials using hybrid exact and fuzzy search of matching eligibility criteria. As a pilot study, patients with non-small cell lung cancer (NSCLC) were recruited from 3 large-volume cancer centres. DNA/RNA CGP were performed using the 523-gene TSO500 panel. ClinMatch recommendations are reviewed in monthly multidisciplinary board meetings.

Results

We curated 72,661 active trials from clinicaltrials.gov and Hong Kong local registries. Among the 193 NSCLC patients recruited between 06/21 and 07/22, 59.7% of patients have at least one clinically actionable alteration. Frequently mutated genes include TP53 (36.0%), EGFR (21.1%) and MET (12.8%). ALK/ROS1/RET fusions were found in 4.7% of patients, while 4.6% of patients have rare fusions such as ETV1, NRG1 and VMP1. KEAP1 mutations were relatively frequent (12.8%), yet contrary to previous report, they are not correlated with tumor mutational burden. Personalised trial suggestions were available for 26.4% of patients.

Conclusions

ClinMatch has shown to be an efficient method to replace labour-intensive clinical trial eligibility criteria matching. Moreover, the platform provides continual updates of trials availability, allowing for retrospective identification of potential subjects who may not have had a suitable trial. This has potential to improve clinical trial accessibility and treatment outcome for patients.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

Codex Genetics Limited.

Funding

Enterprise Support Scheme (S/E004/20), Innovation and Technology Commission, HKSAR; Novartis Pharmaceuticals Corporation; Codex Genetics Limited.

Disclosure

A.C.S. Yu, A.K. Yim, N. Jin, T.C.M. Kan, H.C. Lee: Financial Interests, Institutional, Full or part-time Employment: Codex Genetics Limited. A. Chang, V. Prêtre, A.K.L. Kai: Financial Interests, Institutional, Full or part-time Employment: Novartis Pharmaceuticals Corporation. T.F. Chan: Financial Interests, Institutional, Advisory Role: Codex Genetics Limited. H.H.F. Loong: Financial Interests, Institutional, Invited Speaker: Boehringer Ingelheim, MSD; Financial Interests, Personal, Invited Speaker: Eli-Lilly, Illumina, Bayer, Guardant Health; Financial Interests, Personal, Advisory Board: Novartis, Takeda. All other authors have declared no conflicts of interest.