Multilevel Segmentation of Gleason Patterns using Convolutional Representations in Histopathological Images
Abstract
The Gleason score is the most widely used grading system to diagnose and quantify the aggressiveness of prostate cancer, stratifying regional abnormal patterns on histological images. Nonetheless, recent studies into the Gleason score have reported moderate concordance values of 0.55 (kappa value) in the diagnosis of the disease. This study introduces a convolutional representation for the semantic segmentation and stratification of regions in histological images implementing the Gleason score and three levels of representation. On the first level, a regional network of the Mask R-CNN type is trained with complete annotations to define regional delineations, being effective in locations with general structures. On the second level, using the same architecture, a model is trained only with overlapping annotations from the first scheme, which are difficult-to-classify regions. Finally, a third level of representation produces a more granular description of the regions, considering the regions resulting from the activations of the first level. The final segmentation results from the superposition of the three levels of representation. The proposed strategy was validated and trained on a public set with 886 histological images. The segmentations thus generated achieved an average Area Under the Precision-Recall Curve (AUPRC) of 0.8 ± 0.18 and 0.76 ± 0.15 regarding the diagnoses of two pathologists, respectively. The results show regional intersection levels close to those of the reference pathologists. The proposed strategy is a potential tool to be implemented in clinical support and analysis.
References
F. Bray; J. Ferlay; I. Soerjomataram; R. L. Siegel; L. A. Torre; A. Jemal, “Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries,” CA: A Cancer j. Clinic., vol. 68, no. 6, pp. 394–424, Nov. 2018. https://doi.org/10.3322/caac.21492
E. Bley; A. Silva, “Diagnóstico precoz del cáncer de próstata,” Rev.Méd. Clín. Las Condes, vol. 22, no. 4, pp. 453–458, Jul. 2011. https://doi.org/10.1016/S0716-8640(11)70450-3
A. I. Ruiz López; J. C. Pérez Mesa, Y. Cruz Batista; L. E. González Lorenzo, “Actualización sobre cáncer de próstata,” ccm, vol. 21, no. 3, Jul. 2017. http://scielo.sld.cu/scielo.php?script=sci_arttext&pid=S1560-43812017000300021
American Cancer Society, “Pruebas para diagnosticar y determinar la etapa del cáncer de próstata,” 2019. https://www.cancer.org/es/cancer/cancer-de-prostata/deteccion-diagnostico-clasificacion-por-etapas/como-se-diagnostica.html
D. F. R. Griffiths et al., “A study of Gleason score interpretation in different groups of UK pathologists; techniques for improving reproducibility,” Histopathology, vol. 48, no. 6, pp. 655–662, May. 2006. https://doi.org/10.1111/j.1365-2559.2006.02394.x
K. C. Coard; V. L. Freeman, “Gleason Grading of Prostate Cancer: Level of Concordance Between Pathologists at the University Hospital of the West Indies,” Am. J. Clin. Pathol, vol. 122, no. 3, pp. 373–376, Sep. 2004. https://doi.org/10.1309/MHCY35FJ296CLLC8
M. McLean; J Srigley; D. Banerjee; P. Warde; Y. Hao, “Interobserver variation in prostate cancer gleason scoring: Are there implications for the design of clinical trials and treatment strategies?,” Clin. Oncol., vol. 9, no. 4, pp. 222–225, Jan. 1997. https://doi.org/10.1016/S0936-6555(97)80005-2
S. Doyle; M. Hwang; K. Shah; A. Madabhushi; M. Feldman; J. Tomaszeweski, “Automated grading of prostate cancer using architectural and textural image features,” in 4th IEEE Inter. Symp. Biomed. Imaging: From Nano to Macro, pp. 1284–1287, Arlington 2007. http://doi.org/10.1109/ISBI.2007.357094
S. Doyle; A. Madabhushi; M. Feldman; John Tomaszeweski, “A boosting cascade for automated detection of prostate cancer from digitized histology,” in Medical Image Computing and
Computer-Assisted Intervention, R. Larsen; M. Nielsen; J. Sporring, Eds. 2006. https://doi.org/10.1007/11866763_62
R. Farjam; H. Soltanian-Zadeh; K. Jafari-Khouzani; R. A. Zoroofi, “An image analysis approach for automatic malignancy determination of prostate pathological images,” Cytom. Part B Clin. Cytom., vol. 72B, no. 4, pp. 227–240, Jul. 2007. https://doi.org/10.1002/cyto.b.20162
E. Arvaniti et al., “Automated Gleason grading of prostate cancer tissue microarrays via deep learning,” Scientific reports, vol. 8, Aug. 2018. https://doi.org/10.1038/s41598-018-30535-1
F. León; M. Plazas; F. Martínez, “An inception deep architecture to differentiate close-related Gleason prostate cancer scores,” in 15th Inter. Symp. Med. Info. Proces. Anal., Medellín, 2019. https://doi.org/10.1117/12.2547113
E. Payá Bosch, “Diseño y desarrollo de un sistema automático de segmentación de glándulas histológicas para identificar el cáncer de próstata en una etapa inicial,” Tesis de Maestría, Univ. Pol. Valencia, 2019. http://hdl.handle.net/10251/129307
J. G. García Pardo, “Diseño y desarrollo de un sistema automático de clasificación de estructuras glandulares en imágenes histológicas de próstata,” Tesis de maestría, Univ. Pol. Valencia, 2018. http://hdl.handle.net/10251/109246
W. Bulten et al., “Automated deep- learning system for Gleason grading of prostate cancer using biopsies: A diagnostic study,” Lancet Oncol., vol. 21. no. 2, pp. 233- 242, Feb. 2020. https://doi.org/10.1016/S1470-2045(19)30739-9
I. Nathan et al., “Semantic segmentation for prostate cancer grading by convolutional neural networks,” in Medical Imaging 2018: Digital Pathology, Houston, 2018. https://doi.org/10.1117/12.2293000
K. He; G. Gkioxari; P. Dollár; R. Girshick, “Mask r-cnn,” in Proc. IEEE Inter. Conf. Computer vision, Venice, 2017. https://doi.org/10.1109/ICCV.2017.322
A. Babenko Yandex; V. Lempitsky, “Aggregating local deep features for image retrieval,” in 2015 IEEE Inter, Conf. Comp. Vision (ICCV),). https://doi.org/10.1109/ICCV.2015.150
Y. Kalantidis; C. Mellina; S. Osindero, “Crossdimensional weighting for aggregated deep convolutional features,” in European conference on computer vision, Lecture Notes in Computer Science, vol. 9913. Springer, Cham, 2016. https://doi.org/10.1007/978-3-319-46604-0_48
E. Arvaniti et al., “Replication data for: Automated Gleason grading of prostate cancer tissue microarrays via deep learning,” 2018. https://doi.org/10.7910/DVN/OCYCMP
Q. Zhong et al., “A curated collection of tissue microarray images and clinical outcome data of prostate cancer patients,” Scientific data, vol. 4, 170014, Mar. 2017. https://doi.org/10.1038/sdata.2017.14
L. Tsung-Yi et al., “COCO API - Dataset”, 2020. https://github.com/cocodataset/cocoapi
D. García Seisdedos, “Segmentación de núcleos celulares en imágenes de microscopía ayudados por redes neuronales convolucionales,” Trabajo de grado, Univ. Oberta Catalunya, 2018. http://hdl.handle.net/10609/82134
D. Marín Soto, “Segmentación de células mediante técnicas de Procesamiento Digital de Imágenes para el rastreo de células cancerosas,” Trabajo de grado, Inst. Tec. Costa Rica, 2018. https://hdl.handle.net/2238/10383
P. Henderson; V. Ferrari, “End-to-End Training of Object Class Detectors for Mean Average Precision,” in Lecture Notes in Computer Science, vol. 10115, pp. 198–213, Springer, Cham. 2017. https://doi.org/10.1007/978-3-319-54193-8_13
K. Boyd; K. H. Eng; C. D. Page, “Area under the Precision-Recall Curve: Point Estimates and Confidence Intervals,” in Lecture Notes in Computer Science, vol. 8190, Springer, Berlin, Heidelberg. 2013. https://doi.org/10.1007/978-3-642-40994-3_29
Downloads
Altmetric
