RESUMO
Multiplex imaging technologies allow the characterization of single cells in their cellular environments. Understanding the organization of single cells within their microenvironment and quantifying disease-status related biomarkers is essential for multiplex datasets. Here we proposed SNOWFLAKE, a graph neural network framework pipeline for the prediction of disease-status from combined multiplex cell expression and morphology in human B-cell follicles. We applied SNOWFLAKE to a multiplex dataset related to COVID-19 infection in humans and showed better predictive power of the SNOWFLAKE pipeline compared to other machine learning and deep learning methods. Moreover, we combined morphological features inside graph edge features to utilize attribution methods for extracting disease-relevant motifs from single-cell spatial graphs. The underlying subgraphs were further analyzed and associated with disease status across the dataset. We showed that SNOWFLAKE successfully extracted significant low dimensional embedding from subgraphs with a clear separation between disease status and helped characterize unique cellular interactions in the subgraphs. SNOWFLAKE is a generalizable pipeline for the analysis of multiplex imaging data modality by extracting disease-relevant subgraphs guided by graph-level prediction.
RESUMO
BACKGROUND AND PURPOSE: Despite excellent results, widespread acceptance of the laparoscopic dismembered Anderson-Hynes pyeloplasty (AH) is hampered by its steep learning curve. Laparoscopic nondismembered pyeloplasty techniques, although simpler, have not matched the results of AH. We have been using a technical modification of AH to combine its excellent outcome with technical ease of nondismembered pyeloplasties. We describe the procedure and results of laparoscopic postanastomotic dismemberment (PAD) pyeloplasty for primary ureteropelvic junction (UPJ) obstruction. PATIENTS AND METHODS: PAD technique involves an initial partial division of the dilated pelvis and ureteral spatulation without dismembering the UPJ. Both layers of ureteropelvic anastomosis are completed before dismemberment and pelvic reduction. Forty-one PAD procedures in 40 patients with UPJ obstruction and follow-up of at least 3 months were evaluated. Mean age was 37.2 years (range 2-82 years) with 22 patients younger than 15 years. The UPJ was dependent in 31 and had high insertion in 10 (24.4%). The stenotic segment was long (> or =1.5 cm) in 18 (43.9%). Crossing vessels and secondary calculi were observed in six (14.6%) and seven (17.1%) units. RESULTS: Mean (+/- SD) blood loss, hospital stay, convalescence, and analgesia requirement were 68.1 +/- 37.6 mL, 3.8 +/- 1.1 days, 11.4 +/- 3.9 days, and 204.8 +/- 60.5 mg diclofenac, respectively. The mean operative time was 97.6 +/- 22.1 minutes. There was one intraoperative complication in the form of injury to a renal vein tributary, with no transfusions or conversions. Postoperative complications included pain after stent removal, persistent drainage, and pyelonephritis in 1, 2, and 4 patients, respectively. Mean follow-up was 19.5 months (range 3-58 months), with a success rate of 95.1%. Failures were not attributable to UPJ configuration, length of stenosis, or age. CONCLUSIONS: The PAD technique has several practical advantages with a shorter operative time compared with other historical series of laparoscopic pyeloplasty (LP). It combines the ease of nondismembered LP with the excellent outcome of dismembered techniques.