ESTRO 2023

Session Item

Saturday

May 13

15:15 - 16:15

Stolz 1

Prostate

Chair: , Italy;

Co-Chair: William Kinnaird, United Kingdom

Session Code: 1420

Session Type: Mini-Oral

Track: Clinical

Intraprostatic GTV delineation in 18F-PSMA-PET Images for Patients with Primary PCa using a CNN

Julius Holzschuh, Germany

Presentation Number: MO-0221

Abstract

Abstract Title:

Intraprostatic GTV delineation in 18F-PSMA-PET Images for Patients with Primary PCa using a CNN

Authors:

Julius Holzschuh^1,2,3, Michael Mix⁴, Juri Ruf⁵, Tobias Hölscher⁶, Jörg Kotzerke⁷, Alexis Vrachimis⁸, Harun Ilhan⁹, Simon K.B. Spohn^10,11,12, Tobias Fechter^13,10, Dejan Kostyszyn^2,14, Peter Bronsert¹⁵, Christian Gratzke¹⁶, Radu Grosu¹⁷, Sophia C. Kamran¹⁸, Pedram Heidari¹⁹, Thomas S.C. Ng²⁰, Arda Könik²¹, Anca-Ligia Grosu^22,23, Constantinos Zamboglou^1,24

¹Medical Center - University of Freiburg, Department of Radiation Oncology, Freiburg, Germany; ²German Cancer Consortium (DKTK), Partner Site Freiburg, Freiburg, Germany; ³Karlsruhe Institute of Technology, Faculty of Computer Science, Karlsruhe, Germany; ⁴Medical Center - University of Freiburg, Department of Nuclear Medicine, Freiburg, Germany; ⁵Medical Center - University of Freiburg, Department of Nuclear Medicine , Freiburg, Germany; ⁶Faculty of Medicine and University Hospital Carl Gustav Carus, Technische Universität Dresden, Department of Radiotherapy and Radiation Oncology,, Dresden, Germany; ⁷Dresden University Hospital, Department of Nuclear Medicine, Dresden, Germany; ⁸German Oncology Center - University Hospital of the European University, Department of Nuclear Medicine, Limassol, Cyprus; ⁹University Hospital - Ludwig-Maximilians-Universität, Department of Nuclear Medicine, Munich, Germany; ¹⁰ Medical Center - University of Freiburg, Department of Radiation Oncology, Freiburg, Germany; ¹¹ German Cancer Consortium (DKTK), Partner Site Freiburg , Freiburg, Germany; ¹²Faculty of Medicine - University of Freiburg, Berta-Ottenstein-Programme, Freiburg, Germany; ¹³German Cancer Consortium (DKTK), Partner Site Freiburg , Freiburg, Germany; ¹⁴Medical Center - University of Freiburg, Department of Radiation Oncology , Freiburg, Germany; ¹⁵Medical Center - University of Freiburg, Department of Pathology, Freiburg, Germany; ¹⁶Medical Center - University of Freiburg, Department of Urology, Freiburg, Germany; ¹⁷Technical University of Vienna, Department of Computer Science, Vienna, Austria; ¹⁸Massachusetts General Hospital - Harvard Medical School, Department of Radiation Oncology, Boston, USA; ¹⁹Massachusetts General Hospital - Harvard Medical School, Division of Nuclear Medicine and Molecular Imaging, Department of Radiology, Boston, USA; ²⁰Massachusetts General Hospital - Harvard Medical School, Department of Radiology, Boston, USA; ²¹Dana-Farber Cancer Institute - Harvard Medical School, Department of Imaging - Nuclear Medicine, Boston, USA; ²²Medical Center - University of Freiburg, Department of Radiation Oncology, Freiburg, Germany; ²³ German Cancer Consortium (DKTK), Partner Site Freiburg, Freiburg, Germany; ²⁴German Oncology Center - University Hospital of the European University, Department of Radiation Oncology, Limassol, Cyprus

Show Affiliations

Purpose or Objective

With novel radio therapeutic treatment approaches, like focal dose escalation, for primary prostate cancer (PCa) on the rise, an accurate delineation of gross tumor volume (GTV) in prostate-specific membrane antigen PET (PSMA-PET) becomes more and more important. Manual approaches to this task are time consuming, require a lot of clinical expertise and are observer dependent. To overcome this issue, a fully convolutional neural network (CNN) was trained to automate this task and to provide an efficient solution to this problem.

Material and Methods

A modified 3D U-Net was trained on a dataset consisting of 128 different 18F-1007-PSMA-PET images from three different institutions (Freiburg: n=77, Cyprus: n=32, Munich: n=19). Testing was done on one independent internal cohort (Freiburg: n=19) and two independent external cohorts (Boston: n=9 18F-DCFPyL-PSMA, Dresden: n=14 18F-1007-PSMA). Expert contours were generated in consensus for each patient using a validated technique. CNN predictions were compared to expert contours using Dice similarity coefficient (DSC).
Co-registered whole-mount histology was used for the internal testing cohort to compare sensitivity and specificity of CNN predictions and expert contours.

Results

Median DSCs were Freiburg: 0.82 (IQR: 0.73-0.88), Dresden: 0.71 (IQR: 0.53-0.75) and Boston: 0.80 (IQR: 0.64-0.83) respectively.
CNN sensitivity was slightly higher than manual contours with a median sensitivity of 0.88 (IQR: 0.68-0.97) and 0.85 (IQR: 0.75-0.88), respectively. With a median specificity of 0.83 (IQR: 0.57-0.97) and 0.88 (IQR: 0.69-0.98) for CNN and expert contours, respectively, CNN specificity was slightly lower (p<0.05).

Conclusion

Deep learning methods provide a great way to automate complex tasks in modern medicine. Even partially achieving higher sensitivities than experts while not sacrificing too much specificity, fast and reliable GTV segmentation can be obtained for 18F-1007-PSMA-PET and 18F-DCFPyL-PSMA-PET.

Description:

Evaluation process for specificity and sensitivity for GTVs on Freiburg testing cohort. PCa GTVs are delineated by both CNN and experts separately. Co-registered whole mount histology is used as ground truth.