Prostate cancer malignancy detection and localization from mpMRI using auto-deep learning as one step closer to clinical utilization

Weiwei Zong; Eric Carver; Simeng Zhu; Eric Schaff; Daniel Chapman; Joon Lee; Hassan Bagher-Ebadian; Benjamin Movsas; Winston Wen; Tarik Alafif; Xiangyun Zong

doi:10.1038/s41598-022-27007-y

Prostate cancer malignancy detection and localization from mpMRI using auto-deep learning as one step closer to clinical utilization

Sci Rep. 2022 Dec 27;12(1):22430. doi: 10.1038/s41598-022-27007-y.

Authors

Weiwei Zong¹, Eric Carver², Simeng Zhu², Eric Schaff², Daniel Chapman², Joon Lee³, Hassan Bagher-Ebadian^{2

4

5

6}, Benjamin Movsas², Winston Wen^{7

8

9}, Tarik Alafif¹⁰, Xiangyun Zong¹¹

Affiliations

¹ WeCare.WeTeach, Troy, MI, 48098, USA. mandyzong.research@gmail.com.
² Henry Ford Health System, Detroit, MI, 48202, USA.
³ Trinity Health, Minot, ND, 58701, USA.
⁴ Department of Radiology, Michigan State University, East Lansing, MI, 48824, USA.
⁵ Department of Osteophytic, Michigan State University, East Lansing, MI, 48824, USA.
⁶ Department of Physics, Oakland University, Rochester, MI, 48309, USA.
⁷ SJTU-Ruijing-UIH Institute for Medical Imaging Technology, Shanghai, 200241, China.
⁸ Department of Radiology, Ruijin Hospital Shanghai Jiaotong University School of Medicine, Shanghai, 200031, China.
⁹ The Global Institute of Future Technology, Shanghai Jiaotong University, Shanghai, 200240, China.
¹⁰ Umm Al-Qura University, Jamoum, 25375, Saudi Arabia.
¹¹ Shanghai JiaoTong University Affiliated Sixth People's Hospital, Shanghai, 200233, China.

Abstract

Automatic diagnosis of malignant prostate cancer patients from mpMRI has been studied heavily in the past years. Model interpretation and domain drift have been the main road blocks for clinical utilization. As an extension from our previous work we trained on a public cohort with 201 patients and the cropped 2.5D slices of the prostate glands were used as the input, and the optimal model were searched in the model space using autoKeras. As an innovative move, peripheral zone (PZ) and central gland (CG) were trained and tested separately, the PZ detector and CG detector were demonstrated effective in highlighting the most suspicious slices out of a sequence, hopefully to greatly ease the workload for the physicians.

Publication types

Research Support, Non-U.S. Gov't

MeSH terms

Deep Learning*
Humans
Magnetic Resonance Imaging
Male
Multiparametric Magnetic Resonance Imaging*
Prostate / pathology
Prostatic Neoplasms* / diagnosis
Prostatic Neoplasms* / pathology