Share Email Print
cover

Proceedings Paper • new

Deep learning based bladder cancer treatment response assessment
Author(s): Eric Wu; Lubomir M. Hadjiiski; Ravi K. Samala; Heang-Ping Chan; Kenny H. Cha; Caleb Richter; Richard H. Cohan; Elaine M. Caoili; Chintana Paramagul; Ajjai Alva; Alon Z. Weizer
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

We compared the performance of different Deep Learning - Convolutional Neural Network (DL-CNN) models for bladder cancer treatment response assessment based on transfer learning by freezing different DL-CNN layers and variation of the DL-CNN structure. Pre- and post-treatment CT scans of 123 patients (129 cancers, 158 pre- and posttreatment cancer pairs) undergoing chemotherapy were collected. 33% of patients had T0 stage cancer (complete response) after chemotherapy. Regions of interest (ROIs) of pre- and post-treatment scans were extracted from the segmented lesions and combined into hybrid pre-post image pairs. The dataset was split into training (94 pairs and 6209 hybrid ROIs), validation (10 pairs) and test sets (54 pairs). The DL-CNN consists of 2 convolution (C1, C2), 2 locally connected (L1, L2), and 1 fully connected layers, implemented in TensorFlow. The DL-CNN was trained to classify the bladder cancers as fully responding (stage T0) or not fully responding to chemotherapy based on the hybrid ROIs. Two blinded radiologists provided an estimate of the likelihood of the lesion being stage T0 post-treatment by reading the pairs of pre- and post-treatment CT volumes. The test AUC was 0.73 for T0 prediction by the base DL-CNN structure with randomly initialized weights. The base DL-CNN structure with transfer learning pre-trained weights (no frozen layers) achieved a test AUC of 0.79. The test AUCs for 3 modified DL-CNN structures (different C1, C2 max pooling filter sizes, strides, and padding, with transfer learning) were 0.72, 0.86, and 0.69, respectively. For the base DL-CNN with (C1) frozen, (C1, C2) frozen, and (C1, C2, L3) frozen during transfer learning, the test AUCs were 0.81, 0.78, and 0.71, respectively. The radiologists’ AUCs were 0.76 and 0.77. The DL-CNN performed better with pre-trained than randomly initialized weights.

Paper Details

Date Published: 13 March 2019
PDF: 7 pages
Proc. SPIE 10950, Medical Imaging 2019: Computer-Aided Diagnosis, 109503D (13 March 2019); doi: 10.1117/12.2512240
Show Author Affiliations
Eric Wu, Univ. of Michigan (United States)
Lubomir M. Hadjiiski, Univ. of Michigan (United States)
Ravi K. Samala, Univ. of Michigan (United States)
Heang-Ping Chan, Univ. of Michigan (United States)
Kenny H. Cha, U.S. Food and Drug Administration (United States)
Caleb Richter, Univ. of Michigan (United States)
Richard H. Cohan, Univ. of Michigan (United States)
Elaine M. Caoili, Univ. of Michigan (United States)
Chintana Paramagul, Univ. of Michigan (United States)
Ajjai Alva, Univ. of Michigan (United States)
Alon Z. Weizer, Univ. of Michigan (United States)


Published in SPIE Proceedings Vol. 10950:
Medical Imaging 2019: Computer-Aided Diagnosis
Kensaku Mori; Horst K. Hahn, Editor(s)

© SPIE. Terms of Use
Back to Top