Novel Pixelwise Co-Registered Hematoxylin-Eosin and Multiphoton Microscopy Image Dataset for Human Colon Lesion Diagnosis

The dataset consists of a set of 50 samples of lesions obtained by colonoscopies and colectomies carried out between the years 2012 and 2017 at at Digestive Service at OSI Bilbao Basurto. These are 24 malignant neoplasms (adenocarcinoma), 19 preneoplastic lesions (adenoma) and 2 hyperplasia and 5 healthy tissues, obtained from 24 men and 19 women. The samples were diagnosed by the Pathological Anatomy Department at OSI Bilbao Basurto and the FFPE (Formalin-Fixed Paraffin-Embedded) blocks were stored in the Basque Biobank. All the samples were processed after signing Informed Consent and following standard operation procedures. The samples were scanned using a multiphoton microscope (Lens, Florence, Italy) both for multi-photon fluorescence (MPM) and second harmonic generation (SHG) and later stained with H&E (Hematoxylin & Eosin). The different image modalities were reconstructed and corregistered by performing non-rigid deformation (Tecnalia, Bilbao, Spain) allowing pixel correspondence among the different modalities. Pathologists from Basurto Hospital manually labeled the regions where the lession is present. Scale of current dataset is 0.5um/px. Technical Details: On the data_info.csv information for each sample is included: - SAMPLE_ID: Includes the tag "PICCOLO_XX_YY", where XX stands for the patient ID number and YY to the lesion suffix ID. - LESION_ID: The lesion is classified as healthy, hyperplasia, benign neoplasia, and malign neoplasia. - HISTOLOGICAL ANALYSIS: It includes informative test about the lesion. - GRADE: In the case for malign neoplasia universal grading system grade is included.