Revalidation of PET/computed tomography criteria (Hopkins criteria) for the assessment of therapeutic response in lung cancer patients: inter-reader reliability, accuracy and survival outcomes.

BACKGROUND/AIM: Systematic reporting using qualitative evaluation of PET/computed tomography (CT) results has been demonstrated to be very accurate and reproducible in posttherapy assessment of lung cancer (so-called Hopkins criteria). Our aim was to test, in a different cohort of patients, the Hopkins criteria for assessment of therapeutic response in lung cancer and to compare the results with those obtained using a semi-quantitative evaluation of uptake.

METHODS

This is a retrospective study. A total of 85 patients with known lung cancer who underwent fluorine-18 fluorodeoxyglucose PET/CT assessment within 24 weeks (mean 7.9 weeks) of completion of treatment were included. Treatments included surgical resection, chemotherapy, radiation therapy, immunotherapy or combinations thereof. PET/CT interpretation was done by two nuclear medicine physicians, and discrepancies were resolved by a third interpreter. Studies were scored both according to the Hopkins criteria using qualitative assessment of tracer uptake for the primary tumour, locoregional disease in the mediastinum and distant metastatic sites and by applying the same five-point score using a semi-quantitative measure, maximum standardized uptake value. Overall scores of 1, 2 and 3 were considered negative for residual disease, while scores of 4 and 5 were considered positive. Patients were followed up for a median of 18.5 months (range 2-139 months). Kaplan-Meier plots with a Mantel-Cox log-rank test were performed, considering death as the endpoint. Inter-reader variability was assessed using percent agreement and kappa statistics.

RESULTS

The Cohen κ coefficient analysis showed substantial agreement between the two interpreters on the five-point Hopkins criteria scoring, with a κ of 0.73. There was almost perfect agreement between the interpreters with respect to classification as positive or negative according to the Hopkins criteria, with a κ of 0.89. The sensitivity, specificity, positive predictive value, negative predictive value and accuracy of the Hopkins criteria were 88.5% [95% confidence interval (CI) 80.6-96.5%), 79.2% (95% CI 63.2-95.1%), 91.5% (95% CI 84.4-98.6%), 73.1% (95% CI 61.8-84.4%) and 85.9% (95% CI 78.5-93.3%), respectively. There was almost perfect agreement between the qualitative and semi-quantitative scoring with a κ of 0.87, with sensitivity, specificity, positive predictive value, negative predictive value and accuracy of the semi-quantitative Hopkin's criteria of 86.9% (95% CI 78.4-95.4%), 79.2% (95% CI 62.9-95.4%), 91.4% (95% CI 84.2-98.6%), 70.4% (95% CI 58.6-82.1%) and 84.7% (95% CI 80.8-92.4%), respectively.

CONCLUSION

The use of Hopkins criteria for posttherapy assessment in patients with lung cancer represents an easy and reproducible method with substantial to almost perfect interobserver agreement and high positive predictive value and accuracy; moreover, it is easily understood by referring physicians. Additionally, there was no significant difference when applying a semi-quantitative measure to the same five-point score.