Longitudinal lung cancer prediction convolutional neural network model improves the classification of indeterminate pulmonary nodules

Abstract

A deep learning model (LCP CNN) for the stratification of indeterminate pulmonary nodules (IPNs) demonstrated better discrimination than commonly used clinical prediction models. However, the LCP CNN score is based on a single timepoint that ignores longitudinal information when prior imaging studies are available. Clinically, IPNs are often followed over time and temporal trends in nodule size or morphology inform management. In this study we investigated whether the change in LCP CNN scores over time was different between benign and malignant nodules. This study used a prospective-specimen collection, retrospective-blinded-evaluation (PRoBE) design. Subjects with incidentally or screening detected IPNs 6-30 mm in diameter with at least 3 consecutive CT scans prior to diagnosis (slice thickness ≤ 1.5 mm) with the same nodule present were included. Disease outcome was adjudicated by biopsy-proven malignancy, biopsy-proven benign disease and absence of growth on at least 2-year imaging follow-up. Lung nodules were analyzed using the Optellum LCP CNN model. Investigators performing image analysis were blinded to all clinical data. The LCP CNN score was determined for 48 benign and 32 malignant nodules. There was no significant difference in the initial LCP CNN score between benign and malignant nodules. Overall, the LCP CNN scores of benign nodules remained relatively stable over time while that of malignant nodules continued to increase over time. The difference in these two trends was statistically significant. We also developed a joint model that incorporates longitudinal LCP CNN scores to predict future probability of cancer. Malignant and benign nodules appear to have distinctive trends in LCP CNN score over time. This suggests that longitudinal modeling may improve radiomic prediction of lung cancer over current models. Additional studies are needed to validate these early findings.

Conflict of interest statement

The authors declare no competing interests.

© 2023. The Author(s).

Figures

Figure 1

Spaghetti plot with LCP CNN score over time for benign and malignant nodules.

Figure 2

The predicted LCP CNN score on a square root scale overtime for benign and malignant nodules. Panel A shows the predicted LCP CNN score with time 0 representing the time of initial nodule identification. Panel B shows the predicted LCP CNN score with time 0 representing the time of final nodule diagnosis.

Figure 3

Dynamic predicted probabilities of event-free (non-malignant) survival for a benign patient 6 (top panel) and a malignant patient 78 (bottom panel) estimated with the joint model. The LCP CNN longitudinal score (star points) for the second visit (left panel) and the third visit (right panel) are shown. The vertical dotted lines represent the time point of the last probability. Left of the vertical line, the fitted longitudinal trajectory (solid line) is depicted. Right of the vertical line, prediction of the conditional probabilities of event-free survival (solid line) with 95% confidence intervals (dashed line).

Figure 4

Conditional event-free survival (probability) with 95% confidence intervals for the benign patient 6 (top panels) and the malignant patient 78 (bottom panels) at t + 3 (left panels), t + 12 (middle panels), and t + 24 months (right panels), where t represents follow up visits. In this example, benign patient 6 had six visits and malignant patient 78 had three visits.

Figure 5

Time-dependent AUC under four different longitudinal sub-models in the joint model for time interval (t, t + 3] (next 3 months, left panel), (t, t + 12] (next 12 months, middle panel), and (t, t + 24] (next 24 months, right panel) given the longitudinal LCP CNN scores to t = 3, 6, 12, and 24 months. The sub-model that includes the longitudinal LCP CNN scores plus the change rate performed the best among the four sub-models. Last value—last LCP CNN score, value—longitudinal LCP CNN score, slope—rate of change in LCP CNN score, value + slope—longitudinal LCP CNN score plus rate of change in LCP CNN score.