| 1. |
- Janko, Matthew, et al.
(författare)
-
Contemporary Outcomes After Partial Resection of Infected Aortic Grafts
- 2021
-
Ingår i: Annals of Vascular Surgery. - : Elsevier. - 0890-5096 .- 1615-5947. ; 76, s. 202-210
-
Tidskriftsartikel (refereegranskat)abstract
- INTRODUCTION: Aortic graft infection remains a considerable clinical challenge, and it is unclear which variables are associated with adverse outcomes among patients undergoing partial resection.METHODS: A retrospective, multi-institutional study of patients who underwent partial resection of infected aortic grafts from 2002 to 2014 was performed using a standard database. Baseline demographics, comorbidities, operative, and postoperative variables were recorded. The primary outcome was mortality. Descriptive statistics, Kaplan-Meier (KM) survival analysis, and Cox regression analysis were performed.RESULTS: One hundred fourteen patients at 22 medical centers in 6 countries underwent partial resection of an infected aortic graft. Seventy percent were men with median age 70 years. Ninety-seven percent had a history of open aortic bypass graft: 88 (77%) patients had infected aortobifemoral bypass, 18 (16%) had infected aortobiiliac bypass, and 1 (0.8%) had an infected thoracic graft. Infection was diagnosed at a median 4.3 years post-implant. All patients underwent partial resection followed by either extra-anatomic (47%) or in situ (53%) vascular reconstruction. Median follow-up period was 17 months (IQR 1, 50 months). Thirty-day mortality was 17.5%. The KM-estimated median survival from time of partial resection was 3.6 years. There was no significant survival difference between those undergoing in situ reconstruction or extra-anatomic bypass (P = 0.6). During follow up, 72% of repairs remained patent and 11% of patients underwent major amputation. On univariate Cox regression analysis, Candida infection was associated with increased risk of mortality (HR 2.4; P = 0.01) as well as aortoenteric fistula (HR 1.9, P = 0.03). Resection of a single graft limb only to resection of abdominal (graft main body) infection was associated with decreased risk of mortality (HR 0.57, P = 0.04), as well as those with American Society of Anesthesiologists classification less than 3 (HR 0.35, P = 0.04). Multivariate analysis did not reveal any factors significantly associated with mortality. Persistent early infection was noted in 26% of patients within 30 days postoperatively, and 39% of patients were found to have any post-repair infection during the follow-up period. Two patients (1.8%) were found to have a late reinfection without early persistent postoperative infection. Patients with any post-repair infection were older (67 vs. 60 years, P = 0.01) and less likely to have patent repairs during follow up (59% vs. 32%, P = 0.01). Patients with aortoenteric fistula had a higher rate of any post-repair infection (63% vs. 29%, P < 0.01)CONCLUSION: This large multi-center study suggests that patients who have undergone partial resection of infected aortic grafts may be at high risk of death or post-repair infection, especially older patients with abdominal infection not isolated to a single graft limb, or with Candida infection or aortoenteric fistula. Late reinfection correlated strongly with early persistent postoperative infection, raising concern for occult retained infected graft material.
|
|
| 2. |
- Mehta, Raghav, et al.
(författare)
-
QU-BraTS : MICCAI BraTS 2020 Challenge on QuantifyingUncertainty in Brain Tumor Segmentation - Analysis of Ranking Scores and Benchmarking Results
- 2022
-
Ingår i: Journal of Machine Learning for Biomedical Imaging. - 2766-905X. ; , s. 1-54
-
Tidskriftsartikel (refereegranskat)abstract
- Deep learning (DL) models have provided the state-of-the-art performance in a wide variety of medical imaging benchmarking challenges, including the Brain Tumor Segmentation (BraTS) challenges. However, the task of focal pathology multi-compartment segmentation (e.g., tumor and lesion sub-regions) is particularly challenging, and potential errors hinder the translation of DL models into clinical workflows. Quantifying the reliability of DL model predictions in the form of uncertainties, could enable clinical review of the most uncertain regions, thereby building trust and paving the way towards clinical translation. Recently, a number of uncertainty estimation methods have been introduced for DL medical image segmentation tasks. Developing scores to evaluate and compare the performance of uncertainty measures will assist the end-user in making more informed decisions. In this study, we explore and evaluate a score developed during the BraTS 2019-2020 task on uncertainty quantification (QU-BraTS), and designed to assess and rank uncertainty estimates for brain tumor multi-compartment segmentation. This score (1) rewards uncertainty estimates that produce high confidence in correct assertions, and those that assign low confidence levels at incorrect assertions, and (2) penalizes uncertainty measures that lead to a higher percentages of under-confident correct assertions. We further benchmark the segmentation uncertainties generated by 14 independent participating teams of QU-BraTS 2020, all of which also participated in the main BraTS segmentation task. Overall, our findings confirm the importance and complementary value that uncertainty estimates provide to segmentation algorithms, and hence highlight the need for uncertainty quantification in medical image analyses. Our evaluation code is made publicly available at https://github.com/RagMeh11/QU-BraTS
|
|
