A dataset of tumour-infiltrating lymphocytes in colorectal cancer patients using limited resources.
Anis HasnaouiHelal ImenZouhour Ben AzouzHmidi AmiraRaja JouiniAschraf Chadli-DebbichePublished in: Database : the journal of biological databases and curation (2023)
In the realm of cancer research, specifically focusing on colorectal carcinomas (CRCs), a novel diagnostic test referred to as 'Immunoscore' (IS) has emerged. This test relies on assessing the density of tumour-infiltrating lymphocytes, specifically CD3 and CD8, in both the centre of the tumour (CT) and its invasive margin (IM). IS holds promise as a potential prognostic factor. A retrospective descriptive study was conducted within the Pathology Department of Habib Thameur Hospital in Tunis, Tunisia. The study's aim was to evaluate the prognostic efficacy of IS for patients with CRC by means of a comprehensive survival analysis. This publication introduces the immunoscore in colorectal cancer (ISCRC) dataset, which was meticulously compiled during the aforementioned study. The ISCRC dataset comprises digital slide images sourced from biopsies of 104 patients diagnosed with CRC. Using the tissue microarray technique, an immunohistochemical investigation involving anti-CD3 and anti-CD8 markers was performed in regions designated as 'Hot Spots' within the CT and IM. The images were captured using a smartphone camera. Each marker's percentage presence within its respective region was quantified. The IS was estimated utilizing a semi-quantitative method. The ISCRC dataset encompasses anonymized personal data, along with macroscopic and microscopic attributes. The captured images, acquired through manual efforts using smartphones, stand as a valuable asset for the advancement of predictive algorithms Importantly, the potential applications of these models extend beyond mere prediction capabilities. They lay the groundwork for innovative mobile applications that could potentially revolutionize the practices of pathologists, particularly in healthcare settings constrained by resources and the absence of specialized scanning equipment. Database URL: https://figshare.com/s/5b4fa3e58c247a4851d4.
Keyphrases
- prognostic factors
- healthcare
- deep learning
- convolutional neural network
- computed tomography
- high resolution
- ejection fraction
- machine learning
- end stage renal disease
- optical coherence tomography
- primary care
- big data
- newly diagnosed
- magnetic resonance imaging
- nk cells
- palliative care
- image quality
- human health
- risk assessment
- contrast enhanced
- chronic kidney disease
- dual energy
- squamous cell carcinoma
- mass spectrometry
- papillary thyroid
- data analysis
- artificial intelligence
- quality improvement
- climate change
- patient reported
- positron emission tomography
- health information
- pet ct