115P Spitz tumor classification using artificial intelligence

ESMO Real World Data and Digital Oncology, vol. 10, pp. 100312

Abstract

routinely collected clinical metadata, or are trained on historic lower-resolution datasets.Methods: We developed a multimodal AI model separately integrating last-available and longitudinal LDCT imaging with clinical metadata to predict lung nodule malignancy risk.We specifically evaluated challenging nodules of uncertain malignancy probability on single-timepoint imaging.The model was trained using one of the world's largest LCS trials (SUMMIT: recruiting from 2019), which screened >13,000 individuals.A 3D ResNet-18 processed image data.A TabNet encoder processed clinical metadata.Scan interval and clinical variables were incorporated with image data through feature-wise linear modulation.Performance was evaluated with 5-fold cross-validation using the area under the receiver operating characteristic curve (AUC), sensitivity (Se), and specificity (Sp).Results: Across the hold-out test folds, the single-timepoint model achieved an average AUC of 0.79±0.03(Se: 0.81±0.07;Sp: 0.66±0.09).Incorporating longitudinal imaging data improved the AUC to 0.85±0.03(Se: 0.79±0.05;Sp: 0.76±0.03).Adding scan-interval information further increased the AUC to 0.86±0.02(Se: 0.84±0.07;Sp: 0.73±0.09).The best-performing model, combining longitudinal imaging data, scan interval, and clinical metadata, achieved an AUC of 0.87±0.03(Se: 0.83±0.09;Sp: 0.80±0.05).Conclusions: These results emphasise that longitudinal imaging with scan-interval information provides a clear improvement over single-timepoint models for malignancy prediction, especially in challenging cases.The addition of metadata yields modest gains.The consistently high sensitivity across models complements the often-reported high specificity of radiologists.Integrating such tools into radiology workflows may offer complementary diagnostic value in lung cancer screening.

Authors 10

1. Ruben T. Lucassen

   University Medical Center Utrecht

   Affiliation as printed

   Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands

2. M. Romers

   University Medical Center Utrecht

   Affiliation as printed

   Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands

3. Chiel F Ebbelaar Aachen Department of Dermatology, Leiden University Medical Center

   Leiden University Medical Center

   Affiliation as printed

   Department of Dermatology, Leiden University Medical Center, Leiden, Netherlands

4. Antien L. Mooyaart

   Erasmus MC

   Affiliation as printed

   Department of Pathology, Erasmus Medical Center, Rotterdam, Netherlands

5. L. Wynaendts

   St. Antonius Ziekenhuis

   Affiliation as printed

   Department of Pathology, St. Antonius Hospital, Nieuwegein, Netherlands

6. Nikolas Stathonikos

   University Medical Center Utrecht

   Affiliation as printed

   Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands

7. Gerben Eise Breimer

   University Medical Center Utrecht

   Affiliation as printed

   Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands

8. Anne M.L. Jansen

   University Medical Center Utrecht

   Affiliation as printed

   Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands

9. Mitko Veta

   Eindhoven University of Technology

   Affiliation as printed

   Department of Biomedical Engineering, Eindhoven University of Technology, Eindhoven, Netherlands

10. Willeke A.M. Blokx

    University Medical Center Utrecht

    Affiliation as printed

    Department of Pathology, University Medical Center Utrecht, Utrecht, Netherlands