Clinical urine microscopy for urinary tract infections

Urinary tract infections (UTI) are a common disorder. Its diagnosis can be made by microscopic examination of voided urine for cellular markers of infection. We present a dataset containing 300 images and 3,562 manually annotated urinary cells labelled into seven classes of clinically significant urinary content. It is an enriched dataset with samples acquired from the unstained and untreated urine of patients with symptomatic UTI. The aim of the dataset is to facilitate UTI diagnosis in nearly all clinical settings by using a simple imaging system which leverages advanced machine learning techniques. 

<strong>How to cite us</strong>

Liou, Natasha, Trina De, Adrian Urbanski, Catherine Chieng, Qingyang Kong, Anna L. David, Rajvinder Khasriya, Artur Yakimovich, and Harry Horsley. "A clinical microscopy dataset to develop a deep learning diagnostic test for urinary tract infection." <em>Scientific Data</em> 11, no. 1 (2024): 155.

<pre>@article{liou2024clinical,
title={A clinical microscopy dataset to develop a deep learning diagnostic test for urinary tract infection},
author={Liou, Natasha and De, Trina and Urbanski, Adrian and Chieng, Catherine and Kong, Qingyang and David, Anna L and Khasriya, Rajvinder and Yakimovich, Artur and Horsley, Harry},
journal={Scientific Data},
volume={11},
number={1},
pages={155},
year={2024},
publisher={Nature Publishing Group UK London}
}</pre>

<strong>Download Timeout Troubleshooting</strong>

Use "-C" flag of curl in case you experience timeout of the download:

curl -C - https://rodare...tar.gz_part1\?download\=1 --output ...tar.gz_part1

<strong>Data acquisition </strong>

300 urine samples were obtained from patients with symptomatic UTI between April and August 2022 from a specialist LUTS outpatient clinic in central London. Urine samples were collected as natural voids and processed on-site within one hour to mitigate cellular degradation. Brightfield microscopic examination (Olympus BX41F microscope frame, U-5RE quintuple nosepiece, U-LS30 LED illuminator, U-AC Abbe condenser) was performed at x20 objective (Olympus PLCN20x Plan C N Achromat 20x/0.4). A disposable haemocytometer (C Chip™) was used for enumeration of red cells (RBC), white cells (WBC), epithelial cells (EPC), and the presence of other cellular content per 1 µl of urine by two experienced microscopists.

Images were acquired using the aforementioned brightfield microscope using a 0.5X C-mount adapter connected to a digital colour camera (Infinity 3S-1UR, Teledyne Lumenera). Images were taken in 16-bit colour in 1392 x 1040 .tif format using Capture and Analyse software. An enriched dataset approach was taken to maximise urinary cellular content in the acquired images. Such data curation was also necessary to overcome class imbalance. Daily Kohler illumination and global white balance was performed to ensure consistency in image acquisition. 

<strong>Dataset annotation</strong>

300 images were acquired and manually annotated by first identifying cells of interest as a binary semantic segmentation task. Individual pixels were dichotomously labelled as either informative cells, foreground, or non-informative background. Non-informative background was further constrained by including unidentifiable cells, such as debris or grossly out-of-focus particles. Binary annotation was initially performed using ilastik, an open-source software using a Random Forest classifier for pixel classification, then manually refined at the pixel level to ensure accurate semantic segmentation. This produced a binary mask in 1392 x 1040 .tif format for each corresponding raw colour image. 

Objects of interest were then manually labelled by two expert microscopists into one of seven clinically significant multi-class categories: rods, RBC/WBC, yeast, miscellaneous, single EPC, small EPC sheet, and large EPC sheet. This produced a multi-class mask in 1392 x 1040 .tif format with a label as pixel value from 0-7, where 0 is background (Table 1). 

<strong>Data structure </strong>

The dataset is organised into three root folders: img (image), bin_mask (binary mask), and mult_mask (multi-class mask). Each folder has 300 files in .tif format and labelled with an incremental number.

<strong>Table1</strong>

<pre><code class="language-markdown">Folder Files  Objects  Count Pixel Values

img 300 Raw data 0-255
bin_mask  300 Background/Foreground 0/1
mult_mask  300 Background/Class 0
Rod 1697 1
RBC/WBC 1056 2
Yeast 41 3
Miscellaneous  550 4
Single EPC 182 5
Small EPC sheet 26 6
Large EPC sheet  10 7

Total 3562 </code></pre>

尿路感染（Urinary tract infections, UTI）是一类常见病症。其诊断可通过对自然排尿尿液进行显微镜检查，以检测感染的细胞标志物。本数据集包含300张图像与3562枚经人工标注的尿路细胞，这些细胞被划分为7类具有临床意义的尿液成分。该数据集的样本取自有症状尿路感染患者的未染色、未处理尿液，属于富集型数据集。本数据集旨在通过搭载先进机器学习技术的简易成像系统，助力几乎所有临床场景下的尿路感染诊断。

<strong>引用说明</strong>

Liou, Natasha, Trina De, Adrian Urbanski, Catherine Chieng, Qingyang Kong, Anna L. David, Rajvinder Khasriya, Artur Yakimovich, and Harry Horsley. "A clinical microscopy dataset to develop a deep learning diagnostic test for urinary tract infection." <em>Scientific Data</em> 11, no. 1 (2024): 155.

<pre>@article{liou2024clinical,
title={A clinical microscopy dataset to develop a deep learning diagnostic test for urinary tract infection},
author={Liou, Natasha and De, Trina and Urbanski, Adrian and Chieng, Catherine and Kong, Qingyang and David, Anna L and Khasriya, Rajvinder and Yakimovich, Artur and Horsley, Harry},
journal={Scientific Data},
volume={11},
number={1},
pages={155},
year={2024},
publisher={Nature Publishing Group UK London}
}</pre>

<strong>下载超时故障排查</strong>

若遇到下载超时问题，可使用curl的-C参数：
curl -C - https://rodare...tar.gz_part1?download=1 --output ...tar.gz_part1

<strong>数据采集</strong>

2022年4月至8月，我们从伦敦市中心一家专门的下尿路症状（Lower Urinary Tract Symptoms, LUTS）门诊收集了300份来自有症状尿路感染患者的尿液样本。尿液样本以自然排尿方式收集，并在1小时内现场处理，以减缓细胞降解。

采用明场显微镜检查：使用奥林巴斯BX41F显微镜机架、U-5RE五孔物镜转换器、U-LS30 LED光源、U-AC阿贝聚光镜，搭配20倍物镜（奥林巴斯PLCN20x Plan C N Achromat 20x/0.4）。两名经验丰富的显微镜技师使用一次性血细胞计数板（C Chip™）对每1μL尿液中的红细胞（Red Blood Cell, RBC）、白细胞（White Blood Cell, WBC）、上皮细胞（Epithelial Cell, EPC）及其他细胞成分进行计数。

图像采集采用上述明场显微镜，搭配0.5倍C型转接器连接至彩色数码相机（Infinity 3S-1UR，Teledyne Lumenera）。图像采用Capture and Analyse软件采集，为16位彩色格式，分辨率1392×1040，保存为.tif文件。本数据集采用富集策略，以最大化采集图像中的尿路细胞占比，同时该数据整理策略也有助于解决类别不平衡问题。每日进行科勒照明校准与全局白平衡调整，以保证图像采集的一致性。

<strong>数据集标注</strong>

共采集300张图像，首先通过二分类语义分割任务对目标细胞进行识别，完成初始人工标注：将每个像素二分类标注为有效细胞（前景）或非有效背景。非有效背景还包含无法识别的细胞成分，如碎屑或严重失焦颗粒。初始二分类标注使用开源软件ilastik完成，该软件基于随机森林分类器实现像素分类，随后再逐像素进行人工修正，以确保语义分割的准确性。由此为每张原始彩色图像生成分辨率为1392×1040的.tif格式二值掩码。

随后由两名资深显微镜技师将目标物体手动标注为7类具有临床意义的多分类类别：杆菌、红细胞/白细胞、酵母菌、杂项、单个上皮细胞、小上皮细胞团、大上皮细胞团。由此生成分辨率为1392×1040的.tif格式多分类掩码，像素值范围为0~7，其中0代表背景（见表1）。

<strong>数据集结构</strong>

本数据集包含三个根目录文件夹：img（图像文件夹）、bin_mask（二值掩码文件夹）与mult_mask（多分类掩码文件夹）。每个文件夹均包含300个.tif格式文件，以递增数字命名。

<strong>表1</strong>

<pre><code class="language-markdown">Folder Files Objects Count Pixel Values

img 300 Raw data 0-255
bin_mask 300 Background/Foreground 0/1
mult_mask 300 Background/Class 0
Rod 1697 1
RBC/WBC 1056 2
Yeast 41 3
Miscellaneous 550 4
Single EPC 182 5
Small EPC sheet 26 6
Large EPC sheet 10 7

Total 3562 </code></pre>

搜集汇总

背景与挑战

背景概述

该数据集包含300张临床尿液显微镜图像和3562个手动标注的尿细胞，分为7个临床相关类别，用于尿路感染的机器学习辅助诊断。数据来自有症状患者的未染色尿液样本，包含原始图像和两种类型的标注掩码，采用专业显微镜设备采集并由专家进行标注。