Premium
Semiautomatic registration of digital histopathology images to in vivo MR images in molded and unmolded prostates
Author(s) -
Starobinets Olga,
Guo Richard,
Simko Jeffry P.,
Kuchinsky Kyle,
Kurhanewicz John,
Carroll Peter R.,
Greene Kirsten L.,
Noworolski Susan M.
Publication year - 2014
Publication title -
journal of magnetic resonance imaging
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.563
H-Index - 160
eISSN - 1522-2586
pISSN - 1053-1807
DOI - 10.1002/jmri.24287
Subject(s) - histopathology , prostatectomy , prostate , fixation (population genetics) , prostate cancer , medicine , nuclear medicine , radiology , biomedical engineering , pathology , cancer , population , environmental health
Purpose To evaluate a semiautomatic software‐based method of registering in vivo prostate MR images to digital histopathology images using two approaches: (i) in which the prostates were molded to simulate distortion due to the endorectal imaging coil before fixation, and (ii) in which the prostates were not molded. Materials and Methods T2‐weighted MR images and digitized whole‐mount histopathology images were acquired for 26 patients with biopsy‐confirmed prostate cancer who underwent radical prostatectomy. Ten excised prostates were molded before fixation. A semiautomatic method was used to align MR images to histopathology. Percent overlap between MR and histopathology images, as well as distances between corresponding anatomical landmarks were calculated and used to evaluate the registration technique for molded and unmolded cases. Results The software successfully morphed histology‐based prostate images into corresponding MR images. Percent overlap improved from 80.4 ± 5.8% before morphing to 99.7 ± 0.62% post morphing. Molded prostates had a smaller distance between landmarks (1.91 ± 0.75 mm) versus unmolded (2.34 ± 0.68 mm), P < 0.08. Conclusion Molding a prostate before fixation provided a better alignment of internal structures within the prostate, but this did not reach statistical significance. Software‐based morphing allowed for nearly complete overlap between the pathology slides and the MR images. J. Magn. Reson. Imaging 2014;39:1223–1229 . © 2013 Wiley Periodicals, Inc .