Computerized spermatogenesis staging (CSS) of mouse testis sections via quantitative histomorphological analysis

Abstract

Spermatogenesis in mammals is a cyclic process of spermatogenic cell development in the seminiferous epithelium that can be subdivided into 12 subsequent stages. Histological staging analysis of testis sections, specifically of seminiferous tubule cross-sections, is the only effective method to evaluate the quality of the spermatogenic process and to determine developmental defects leading to infertility. Such staging analysis, however, is tedious and time-consuming, and it may take a long time to become proficient. We now have developed a Computerized Staging system of Spermatogenesis (CSS) for mouse testis sections through learning of an expert with decades of experience in mouse testis staging. The development of the CSS system comprised three major parts: 1) Developing computational image analysis models for mouse testis sections; 2) Automated classification of each seminiferous tubule cross-section into three stage groups: Early Stages (ES: stages I-V), Middle Stages (MS: stages VI-VIII), and Late Stages (LS: stages IV-XII); 3) Automated classification of MS into distinct stages VI, VII-mVIII, and late VIII based on newly developed histomorphological features. A cohort of 40 H&E stained normal mouse testis sections was built according to three modules where 28 cross-sections were leveraged for developing tubule region segmentation, spermatogenic cells types and multi-concentric-layers segmentation models. The rest of 12 testis cross-sections, approximately 2314 tubules whose stages were manually annotated by two expert testis histologists, served as the basis for developing the CSS system. The CSS system's accuracy of mean and standard deviation (MSD) in identifying ES, MS, and LS were 0.93 ± 0.03, 0.94 ± 0.11, and 0.89 ± 0.05 and 0.85 ± 0.12, 0.88 ± 0.07, and 0.96 ± 0.04 for one with 5 years of experience, respectively. The CSS system's accuracy of MSD in identifying stages VI, VII-mVIII, and late VIII are 0.74 ± 0.03, 0.85 ± 0.04, and 0.78 ± 0.06 and 0.34 ± 0.18, 0.78 ± 0.16, and 0.44 ± 0.25 for one with 5 years of experience, respectively. In terms of time it takes to collect these data, it takes on average 3 hours for a histologist and 1.87 hours for the CSS system to finish evaluating an entire testis section (computed with a PC (I7-6800k 4.0 GHzwith 32GB of RAM & 256G SSD) and a Titan 1080Ti GPU). Therefore, the CSS system is more accurate and faster compared to a human histologist in staging, and further optimization and development will not only lead to a complete staging of all 12 stages of mouse spermatogenesis but also could aid in the future diagnosis of human infertility. Moreover, the top-ranking histomorphological features identified by the CSS classifier are consistent with the primary features used by histologists in discriminating stages VI, VII-mVIII, and late VIII.

Keywords: Computerized staging of spermatogenesis; Deep learning; Mouse testicular section images; Mouse testis histology; Seminiferous tubules; Sperm development; Spermatogenic cell segmentation.

Fig. 1.

Illustration of the stages of spermatogenesis in the mouse (Adapted from Russell et al. (1993)). Early (I-V), middle (VI-VIII), and late (IX-XII) stages of seminiferous tubule epithelial cycle were defined in this paper. Various types of developing spermatogonial cells (In and B), spermatocytes (PI, L, Z, P, D, and m2), and various steps of developing spermatids from 1 to 16. Spermatid development is divided into sixteen steps, so 1 to 16 represent each of sixteen steps of developing spermatids.

Fig. 2.

H&E stained slide of mouse testicular sections at different resolutions and a cross-section of a tubule with types of spermatogenic cells and multi-concentric-layers at 400X magnification. Images at different magnifications of H&E stained mouse testicular cross-sections from a slide image to a tubule section with three main types of spermatogenic cells and three concentric-layers. (a) a digitalized Whole Slide Image (WSI) at 100X magnification, (b) a testicular cross-section at 200X magnification, (c) a tubule cross-section from (b) in Stage VII at 400X magnification showing three concentric layers of spermatogenic cells, round spermatids (purple), spermatogonia and spermatocyte regions (red). The middle of a tubule (c) is the lumen with elongated spermatids and/or sperm in green. Between tubules are the background regions (g) in white. Distinct morphological differences in staining and texture between round spermatid (d), spermatocyte (e), and spermatogonium (f) are highlighted. The manually annotated mask (h) of three types of spermatogenic cells from (c).

Fig. 3.

Diagram describes the datasets and modules in the paper and the connection between three datasets and modules in the paper.

Fig. 4.

The diagram shows how the basis for distinguishing developmental stages I-XII is manually generated (b) by histologist DR in a mouse testicular section (a). The annotated tubule regions (d) and boundaries (e) shown in false-color (f) for (a) were generated based on the result of automated seminiferous tubule segmentation (c) and manual annotation (b) by histologist DR. For stage-identifying criteria see Table 2.

Fig. 5.

The diagram shows the flowchart of the CSS system with three modules for identifying (d) three initial stage groups and (r) stages VI, VII-mVIII, and late VIII in a mouse testicular section (a).

Fig. 6.

The distribution of tubules in 6 mouse testis sections of (a) training and of (b) 6 testing sets.

Fig. 7.

The flowchart of seminiferous tubule segmentation. It comprises training (a-e) and seminiferous tubule segmentation (f-l) phases.

Fig. 8.

The flowchart of types of spermatogenic cells and multi-concentric-layers segmentation.

Fig. 9.

Derivative-of-Gaussian (DtG) with light line filter response for elongated spermatid orientation extraction.

Fig. 10.

Quantitative histomorphometry feature map illustration of a tubule (a). The first and second rows show cell-level (b-g) and region-level (h-n) features, respectively. Cell-level features comprise nuclear shape (b,c), nuclear orientation (d,e), and nuclear texture (f,g). Region-level features comprise region shape (h,i), elongated spermatids orientation (j,k), and region textures (l,m,n).

Fig. 11.

The flowchart of seminiferous tubule classification for Modules 2 and 3. It comprises training (a-d) and seminiferous tubule classification (e-g) phases.

Fig. 12.

Illustration of the tubule region segmentation in a mouse testis section (a) by gLoG Xu et al. (2016a) (b), CoNNACaeF Xu et al. (2019a) (c), and the model presented in this paper (d).

Fig. 13.

Qualitative and quantitative results on a mouse testis section for classification of 1)tubules into 3 initial stage groups: ES, MS, and LS by CSS system (b,c) and histologist HL (d,e) compared to the basis (a) by histologist DR where tubules delineated with red, green, and blue curves represent tubules in ES, MS, and LS, respectively, and 2) MS tubules into stages VI, VII-mVIII, and late VIII by the CSS system (g,h), histologist HL (i,j), and ResNet model (k,l) compared to the basis (f) by histologist DR where tubules delineated with purple, orange, and light blue curves represent tubules in stages VI, VII-mVIII, and late VIII, respectively. Qualitative classification results are shown in (b,d,g,i,k) where the tubules marked with sign “x” represent a different classification compared to the basis. The quantitative results are shown in confusion matrices (c,e,h,j,l) where each row of the matrix represents stages predicted by the CSS classifier(c,h), histologist HL (e,j), and a ResNet (l) classifier for the prediction of the developmental stages, respectively, while each column represents the instances stages annotated by histologist DR. The numbers in parentheses are the number of tubules being classified into relevant stages.

Fig. 14.

The mean and standard deviation of CSS (a,c) and human histologist (b,d) in discriminating ES, MS, and LS (a,b) and stages V, VII-mVIII, and VIII (c,d) across 6 mouse testis sections. (e) is the mean and standard deviation of Resnet in classifying stages V, VII-mVIII, and VIII across 6 mouse testis sections. In confusion matrices (a,b,c,d,e), each row of the matrix represents stages predicted by the CSS classifier(a,c), histologist HL (b,d), and a ResNet (e) classifier for the prediction of the developmental stages, respectively, while each column represents the instances stages annotated by histologist DR.