Highly Efficient Follicular Segmentation in Thyroid Cytopathological Whole Slide Image

Siyan Tao, Yao Guo, Chuang Zhu, Huang Chen, Yue Zhang, Jie Yang, Jun Liu

Abstract

In this paper, we propose a novel method for highly efficient follicular segmentation of thyroid cytopathological Whole Slide Images (WSIs). Firstly, we propose a hybrid segmentation architecture, which integrates a classifier into DeepLab V3 by adding a branch. A large amount of the WSI segmentation time is saved by skipping the irrelevant areas using the classification branch. Secondly, we merge the low-scale fine features into the original atrous spatial pyramid pooling (ASPP) in DeepLab V3 to accurately represent the details in cytopathological images. Thirdly, our hybrid model is trained by a criterion-oriented adaptive loss function, which leads to much faster convergence. Experimental results on a collection of thyroid patches demonstrate that the proposed model reaches 80.9% segmentation accuracy. Furthermore, WSI segmentation time is reduced by 93% using our proposed method, while achieving a WSI-level accuracy of 53.4%.

Introduction

Thyroid cancer diagnosis often relies on Fine Needle Aspiration (FNA), where clinicians examine tissue slides under a microscope. To automate this process, "Whole Slide Images" (WSIs) are utilized. However, these digital slides are massive—often exceeding 210,000 x 140,000 pixels. Processing these images presents two major hurdles:

  1. Computational Cost: The sheer size makes it impossible to load a full slide into memory, and processing every single pixel is computationally prohibitive.
  2. Data Imbalance: In a typical thyroid slide, less than 10% of the area actually contains the follicular cells necessary for diagnosis. The rest is "junk" data—either blank space or "colloid" (a gelatinous substance) that provides no diagnostic value.
Examples of different tissue types found in thyroid WSIs, including follicular clusters and background colloid.

Proposed Methodology

The proposed approach makes segmentation both accurate and efficient through three primary innovations:

Hybrid Architecture

Instead of running a heavy segmentation model over the entire slide, the authors developed a hybrid model based on ResNet-101. It features a classification branch that shares the first block of layers with the segmentation model. This branch acts as a gatekeeper, quickly identifying if a patch is "Follicular," "Colloid," or "Non-info." If a patch is not follicular, the model skips the expensive segmentation step entirely.

Enhanced ASPP (E-ASPP)

Standard DeepLab V3 uses Atrous Spatial Pyramid Pooling (ASPP) to capture multi-scale information but often loses fine details. The authors introduced "Enhanced ASPP," which merges low-scale, high-resolution features from earlier layers (Block 3) back into the final pooling stage to better represent tiny cellular details essential for cytopathology.

The proposed hybrid architecture (a) showing the classification branch (c) and the residual block structure (b).

Criterion-Oriented Adaptive Loss Function

To accelerate training, a novel loss function was introduced that weights the standard cross-entropy loss by a performance criterion (such as Mean Intersection over Union). This forces the model to converge faster by focusing on the metrics that directly impact evaluation performance.

Experimental Setup

The study utilized a private clinical dataset provided by a top-level comprehensive hospital, consisting of 15 WSIs collected from real patients and stained using the Papanicolaou (Pap) method.

  • Data Split: 13 WSIs were used to generate a patch dataset for training and preliminary testing, while 2 WSIs were reserved for final WSI-level performance evaluation.
  • Preprocessing: A color adjustment method was applied to normalize staining variations between different pathology slides, ensuring model robustness.

Key Results and Discussion

  • Efficiency Gains: The most significant result is the 93% reduction in time required for WSI segmentation. While standard models like FCN or DeepLab V3 required nearly 90 minutes to process a single slide, the proposed hybrid model finished in approximately 12.6 minutes.
  • Segmentation Accuracy: On individual patches, the model achieved 80.9% mIoU, outperforming both FCN (80.2%) and standard DeepLab V3 (72.4%).
  • Classification Reliability: The integrated classifier reached 98% accuracy, ensuring that very few follicular areas were accidentally skipped during the high-speed filtering process.
  • WSI-Level Performance: Even on complex, full-sized slides, the model maintained a 53.4% mIoU, which is superior to existing models when applied to real-world clinical "Whole Slide" scenarios.

Conclusion

By integrating a classification-based "skip" strategy and enhancing feature pooling, this research demonstrates that automated thyroid cancer diagnosis can be made significantly more efficient without sacrificing accuracy. While a performance gap remains between patch-level and WSI-level segmentation, the 93% reduction in processing time marks a critical step toward making AI-assisted pathology feasible in high-volume clinical environments.

Summary updated

Abstract

In this paper, we propose a novel method for highly efficient follicular segmentation of thyroid cytopathological WSIs. Firstly, we propose a hybrid segmentation architecture, which integrates a classifier into Deeplab V3 by adding a branch. A large amount of the WSI segmentation time is saved by skipping the irrelevant areas using the classification branch. Secondly, we merge the low scale fine features into the original atrous spatial pyramid pooling (ASPP) in Deeplab V3 to accurately represent the details in cytopathological images. Thirdly, our hybrid model is trained by a criterion-oriented adaptive loss function, which leads the model converging much faster. Experimental results on a collection of thyroid patches demonstrate that the proposed model reaches 80.9% on the segmentation accuracy. Besides, 93% time is reduced for the WSI segmentation by using our proposed method, and the WSI-level accuracy achieves 53.4%.
BibTeX
@inproceedings{tao2020highly,
  author = {Siyan Tao and Yao Guo and Chuang Zhu and Huang Chen and Yue Zhang and Jie Yang and Jun Liu},
  title = {{Highly Efficient Follicular Segmentation in Thyroid Cytopathological Whole Slide Image}},
  booktitle = {Precision Health and Medicine},
  year = {2020},
  url = {https://openreview.net/forum?id=rosZx9v-h-x},
}
Figure