Anomaly detection in mixed high-dimensional molecular data.

MOTIVATION: Mixed molecular data combines continuous and categorical features of the same samples, such as OMICS profiles with genotypes, diagnoses, or patient sex. Like all high-dimensional molecular data, it is prone to incorrect values that can stem from various sources for example the technical limitations of the measurement devices, errors in the sample preparation, or contamination. Most anomaly detection algorithms identify complete samples as outliers or anomalies. However, in most cases, not all measurements of those samples are erroneous but only a few one-dimensional features within the samples are incorrect. These one-dimensional data errors are continuous measurements that are either located outside or inside the normal ranges of their features but in both cases show atypical values given all other continuous and categorical features in the sample. Additionally, categorical anomalies can occur for example when the genotype or diagnosis was submitted wrongly. RESULTS: We introduce ADMIRE (Anomaly Detection using MIxed gRaphical modEls), a novel approach for the detection and correction of anomalies in mixed high-dimensional data. Hereby, we focus on the detection of single (one-dimensional) data errors in the categorical and continuous features of a sample. For that the joint distribution of continuous and categorical features is learned by mixed graphical models, anomalies are detected by the difference between measured and model-based estimations and are corrected using imputation. We evaluated ADMIRE in simulation and by screening for anomalies in one of our own metabolic datasets. In simulation experiments, ADMIRE outperformed the state-of-the-art methods of Local Outlier Factor, stray, and Isolation Forest. AVAILABILITY AND IMPLEMENTATION: All data and code is available at https://github.com/spang-lab/adadmire. ADMIRE is implemented in a Python package called adadmire which can be found at https://pypi.org/project/adadmire.

Long-term quality of life after resection of retroperitoneal soft tissue sarcoma.

INTRODUCTION: Retroperitoneal soft tissue sarcoma (RPS) is characterized by high recurrence rates. Since complete tumor resection, often necessitating multivisceral resection, enables long-term survival in both primary and recurrent disease, health related quality of life (QoL) after RPS resection has attracted increasing interest. However, data regarding this topic is limited. Here, we multidimensionally assessed long-term QoL after RPS resection. METHODS: Five previously validated (1. EORTC QLQ-C30, 2. WEMWBS, 3. FoP-Q-SF, 4. PC-PTSD, 5. Pro-CTCAE) were sent to patients having undergone resection of primary, recurrent and metastasized RPS at Heidelberg University Hospital between 10/2001 and 12/2020. Multivariable linear regression models were used to test associations between clinical/demographic variables and patient reported outcomes (PROs). RESULTS: Questionnaires were answered by 127 patients (71% response rate). The median interval between RPS diagnosis and assessment of PROs was 80 months. The overall Global Health score was 64.1 and comparable to the general German population. RPS patients reported deficits regarding emotional and social functioning, whereas physical limitations were less pronounced. Besides diarrhea, abdominal symptoms were comparable to the overall population. Tumor recurrences, the number of surgeries, multivisceral resections or postoperative complications did not significantly affect long-term QoL ratings. CONCLUSION: RPS patients rate their QoL relatively high, even after multiple and multivisceral resections. Psychosocial well-being should be monitored in follow-up sessions to offer tailored support if necessary, thus improving postoperative care.

Derivatives of prostate-specific antigen as predictors of incidental prostate cancer.

OBJECTIVE To search for an optimal derivative of prostate-specific antigen (PSA) identifying patients at risk of incidental prostate cancer. PATIENTS AND METHODS In all, 693 patients who underwent transurethral resection of the prostate (TURP) with a normal digital rectal examination, no history of prostate cancer and a PSA level of 2.5-10 ng/mL were studied. The total PSA (tPSA), percentage of free/total PSA (%fPSA), complexed PSA (cPSA), PSA density, cPSA density and the ratio of fPSA to cPSA were measured. Specificity, sensitivity, positive and negative predictive values were determined for all possible threshold values indicating the risk of incidental prostate cancer (T1a or T1b). Furthermore, the patients were subdivided according to age and the presence of an indwelling transurethral catheter. The areas under the receiver operating characteristic curves (AUC) were compared. RESULTS In the whole sample, the %fPSA was the best test predicting all incidental prostate cancers (AUC 0.618, reference: tPSA 0.494), whereas cPSA density was the best predictor of T1b disease (AUC 0.720, reference: tPSA 0.548). Stratification by age did not meaningfully alter the results, the presence of a transurethral catheter, however, was associated with a superiority of tests based on fPSA (AUC 0.620-0.670) compared with tests based on tPSA or cPSA (AUC 0.421-0.581). CONCLUSION Replacing tPSA by PSA derivatives (%fPSA or cPSA density) and stratifying by the presence of an indwelling transurethral catheter may improve the prediction of the risk of incidental prostate cancer and spare unnecessary biopsies before TURP in the tPSA range 2.5-10 ng/mL.