Oral Presentation Clinical Oncology Society of Australia Annual Scientific Meeting 2022

RECALL:Reliability of Estimates of Clinicians’Lung cancer caseLoad – a comparison of clinician estimates of key study eligibility criteria with data from the INHALE registry (#52)

Emmy Sneath 1 , Catherine Dunn 1 , Ben Markman 1 2 , Azim Jalali 1 3 , Dishan Herath 3 , Suzanne Kosmider 3 , Ben Brady 2 , Gary Richardson 2 , Phillip Parente 4 , Rachel Wong 4 5 , Anis Hamid 4 , Peter Gibbs 1 3
  1. Personalised Oncology Division, The Walter and Eliza Hall Institute of Medical Research, Parkville, Victoria, Australia
  2. Melbourne Oncology Group, Cabrini Health, Malvern, Victoria, Australia
  3. Department of Medical Oncology, Western Health, Footscray, Victoria, Australia
  4. Department of Medial Oncology, Eastern Health, Box Hill, Victoria, Australia
  5. Monash University, Eastern Health Clinical School, Clayton, Victoria, Australia

Background

Clinical trials are pivotal to progress in cancer therapy. Trial recruitment is frequently slow, with many sites failing to recruit as expected. Ultimately a significant proportion of trials are closed early. One potential explanation is that clinicians’ estimates of the trial eligible patient caseload are unreliable. The INHALE registry collects data for consecutive patients with lung cancer at multiple Victorian sites.The RECALL project compared clinician estimates of patient numbers and key study eligibility criteria with INHALE data.

 
Methods

Medical oncologists at three Victorian INHALE registry sites were invited to complete a survey regarding lung cancer caseload, tumour stage at diagnosis, cancer subtype, molecular data (determining eligibility for a targeted therapy trial), performance status and co-morbidity. Clinician responses were compared with INHALE data for the matching site. Clinician accuracy is  described by dividing the estimate by the site registry number as a percentage.

 

Results

Nine out of 11 clinicians responded. Data on 101 patients was available. Clinicians consistently overestimated site caseload, including for non-small cell (overestimated by 71%), small cell (overestimated by 69%) and mesothelioma (overestimated by 82%). Estimates of both the rate of molecular testing and the proportion of molecular subtypes (adenocarcinoma only) were largely accurate, with the exception of some sites overestimating rates of EGFR mutation or PDL1 >50%. Based on performance status data and estimates of significant co-morbidity and/or poor major organ function data as derived from the modified Charlson index, 71% of patients overall would likely be trial eligible (clinicians estimated 37%).

 

Conclusion

Clinician estimates of patient caseload and proportion meeting key study entry criteria were consistently inaccurate, which could lead to a substantial overestimate of likely trial accruals. A registry based data query is an alternative approach that would likely provide more accurate site estimates, improving clinical trial efficiency.