91 patients with varus knees scheduled for total knee arthroplasty were recruited after getting their informed consent. Inclusion criteria required patients with a diagnosis of idiopathic OA (primarily medial compartment and/or patellofemoral disease) exhibiting a relatively preserved lateral compartment (JSW: 2–10 mm, median: 6 mm in the lateral compartment) based on preoperative weight-bearing anterior-posterior (AP) radio- graphs taken in full extension and 30° of flexion. During TKA, the Lateral Femoral Condyle (LFC) was collected and the AP orientation was noted. All included LFC specimens presented with Grade 0 , I , or II macroscopic Outerbridge classifications. Two osteochondral specimens (4 × 4 × 8 mm) were obtained from each LFC by placing the condyle in an in-house fabricated miter box in AP orientation and cartilage arches were cut using a razor blade from the weight-bearing center portion of the LFC; 1 was located medial and 1 lateral to the LFC midline. The centers of these two samples were separated by 10 mm.

A total of 182 osteochondral specimens were processed from 91 patients. Immediately after surgical retrieval, specimens were fixed for 48 h at 4 °C. 5µ thick paraffin sections were cut and stained with freshly prepared HE or SafraninO and fast green (SafO). Two adjacent sections per stain were digitally imaged under brightfield at 5x and used for scoring using HHGS and advanced OARSI systems. The HHGS and OARSI scores were provided by 3 reviewers, 3 times (separated by at least 3 months). Unstained sections were digitally imaged under polarized light microscopy (PLM) at 1.25X and scored using PLM scoring system.

In addition to the digital high quality tiff images of SafO, HE and PLM sections of cartilage and the respective scores, metadata would also include patient's age, gender, surgery-side.