Automated classification of depression from structural brain measures across two independent community‐based cohorts

Major depressive disorder (MDD) has been the subject of many neuroimaging case–control classification studies. Although some studies report accuracies ≥80%, most have investigated relatively small samples of clinically‐ascertained, currently symptomatic cases, and did not attempt replication in larger samples. We here first aimed to replicate previously reported classification accuracies in a small, well‐phenotyped community‐based group of current MDD cases with clinical interview‐based diagnoses (from STratifying Resilience and Depression Longitudinally cohort, ‘STRADL’). We performed a set of exploratory predictive classification analyses with measures related to brain morphometry and white matter integrity. We applied three classifier types—SVM, penalised logistic regression or decision tree—either with or without optimisation, and with or without feature selection. We then determined whether similar accuracies could be replicated in a larger independent population‐based sample with self‐reported current depression (UK Biobank cohort). Additional analyses extended to lifetime MDD diagnoses—remitted MDD in STRADL, and lifetime‐experienced MDD in UK Biobank. The highest cross‐validation accuracy (75%) was achieved in the initial current MDD sample with a decision tree classifier and cortical surface area features. The most frequently selected decision tree split variables included surface areas of bilateral caudal anterior cingulate, left lingual gyrus, left superior frontal, right precentral and paracentral regions. High accuracy was not achieved in the larger samples with self‐reported current depression (53.73%), with remitted MDD (57.48%), or with lifetime‐experienced MDD (52.68–60.29%). Our results indicate that high predictive classification accuracies may not immediately translate to larger samples with broader criteria for depression, and may not be robust across different classification approaches.     

Use of a Checklist During Observation of a Simulated Cardiac Arrest Scenario Does Not Improve Time to CPR and Defibrillation Over Observation Alone for Subsequent Scenarios

Theory: Immersive simulation is a common mode of education for medical students. Observation of clinical simulations prior to participation is believed to be beneficial, though this is often a passive process. Active observation may be more beneficial. Hypotheses: The hypothesis tested in this study was that the active use of a simple checklist during observation of an immersive simulation would result in better participant performance in a subsequent scenario compared with passive observation alone. Methods: Medical students were randomized to either passive or active (with checklist) observation of an immersive simulation involving cardiac arrest prior to participating in their own simulation. Performance measures included time to cardiopulmonary resuscitation (CPR) and time to defibrillation and were compared between first and second scenarios as well as between passive and active observers. Results: Seventy-nine simulations involving 232 students were conducted. Mean time to CPR was 18 seconds (SD = 11.6) for those using the checklist and 24 seconds (SD = 15.8) for those who observed passively (M difference = 6 seconds), t(35) = 1.46, p =.153. Time to defibrillation was 94 seconds (SD = 26.4) for those using the checklist and 92 seconds (SD = 23.8) for those who observed passively (M difference = –2 seconds), t(38) =.21, p =.837. Time to CPR was 24 seconds (SD = 15.8) for passive observers and 31 seconds (SD = 21.0; M difference = 7 seconds), t(35) = 1.13, p =.265, for their first scenario counterparts. Time to CPR was 18 seconds (SD = 11.6) for active observers and 36 seconds (SD = 26.2; M difference = 18 seconds), t(24) = 2.81, p =.010, for their first scenario counterparts. Time to defibrillation was 92 seconds (SD = 23.8) for passive observers and 125 seconds (SD = 32.2; M difference = 33 seconds), t(33) = 3.63, p =.001, for their first scenario counterparts. Time to defibrillation was 94 seconds (SD = 26.4) for the active observers and 132 seconds (SD = 52.9; M difference = 38 seconds), t(28) =.46, p =.008, for their first scenario counterparts. Conclusions: Observation alone leads to improved performance in the management of a simulated cardiac arrest. The active use of a simple skills-based checklist during observation did not appear to improve performance over passive observation alone.     

Co-targeting bromodomain and extra-terminal proteins and MCL1 induces synergistic cell death in melanoma

The treatment of melanoma has been markedly improved by the introduction of targeted therapies and checkpoint blockade immunotherapy. Unfortunately, resistance to these therapies remains a limitation. Novel anticancer therapeutics targeting the MCL1 anti-apoptotic protein have shown impressive responses in haematological cancers but are yet to be evaluated in melanoma. To assess the sensitivity of melanoma to new MCL1 inhibitors, we measured the response of 51 melanoma cell lines to the novel MCL1 inhibitor, S63845. Additionally, we assessed combination of this drug with inhibitors of the bromodomain and extra-terminal (BET) protein family of epigenetic readers, which we postulated would assist MCL1 inhibition by downregulating anti-apoptotic targets regulated by NF-kB such as BCLXL, BCL2A1 and XIAP, and by upregulating pro-apoptotic proteins including BIM and NOXA. Only 14% of melanoma cell lines showed sensitivity to S63845, however, combination of S63845 and I-BET151 induced highly synergistic apoptotic cell death in all melanoma lines tested and in an in vivo xenograft model. Cell death was dependent on caspases and BAX/BAK. Although the combination of drugs increased the BH3-only protein, BIM, and downregulated anti-apoptotic proteins such as BCL2A1, the importance of these proteins in inducing cell death varied between cell lines. ABT-199 or ABT-263 inhibitors against BCL2 or BCL2 and BCLXL, respectively, induced further cell death when combined with S63845 and I-BET151. The combination of MCL1 and BET inhibition appears to be a promising therapeutic approach for metastatic melanoma, and presents opportunities to add further BCL2 family inhibitors to overcome treatment resistance.     

Outcomes at 3 years of a prospective pilot study of Campath-1H and sirolimus immunosuppression for renal transplantation

Campath-1H (alemtuzumab) induction was used for renal transplantation in combination with sirolimus as immunosuppression. We previously reported a high (28%) rate of early rejection with this regimen, and now report 3-year outcomes. Twenty-nine patients were recipients of either deceased donor or non-HLA (Human Leukocyte Antigen) identical living donor primary renal allografts. Clinical parameters including infection, malignancy, kidney function, and kidney histology were followed prospectively for 3 years. Three-year cumulative graft and patient survival were 96% and 100%, respectively. Twenty patients were maintained on steroid-free immunosuppressive regimens, and 15 patients were maintained on monotherapy for immunosuppression (12 on sirolimus). No serious infectious complications were observed and two patients developed basal cell skin cancer. The 3-year results of our initial pilot study demonstrate good graft (96%) and patient (100%) outcomes. Campath-1H induction has yielded a high proportion of patients maintained on immunosuppressive monotherapy (57%) without serious infectious- and no malignancy-related complications. The reported regimen yielded novel insights into both Campath-1H and sirolimus therapy in renal transplantation. Because of the higher incidence of early rejection, we recommend a modified strategy of immunosuppression including a brief course of a calcineurin inhibitor.     

Relationships Between Arousal-Related Moods and Episodic Tension-Type Headache: A Biopsychological Study

An exploratory study was conducted examining arousal-related moods and episodic tension-type headache. Twelve subjects meeting International Headache Society criteria for episodic tension-type headache and 12 headache-free controls recorded headache activity and mood eight times daily for 14 consecutive days. Moods were measured using the Activation-Deactivation Adjective Check List, a self-report list that subjectively represents general arousal along two dimensions of Tension and Energy. Headache subjects had higher Tension levels than controls even in the absence of pain, and greater variation in this dimension as well. Within the headache group, Tension during pain-free periods was significantly lower than when experiencing headache, and was correlated with headache activity. The results were taken to support Thayer's (1989) biopsychological model of mood and arousal, and are discussed in terms of the model's heuristic value for general arousal and headache research.     

Effects of acute administration of d-amphetamine and haloperidol on procedural learning in man

The effects of an indirect dopamine-agonist, d-amphetamine, and a non-selective dopamine receptor antagonist, haloperidol, were investigated in normal male volunteers using a between-subjects double-blind design in a procedural learning task, thought mainly to involve unconscious/automatic learning. The results showed: (1) d-amphetamine facilitated response speed, whereas haloperidol inhibited it, in comparison to placebo; (2) the linear increase in procedural learning corresponded with pharmacological manipulation of degree of dopaminergic activity, i.e. subjects given haloperidol showed the least, and subjects given d-amphetamine the greatest, procedural learning. The implications of these findings are discussed in relation to investigation of abnormalities of procedural learning processes in schizophrenia. Received: 28 June 1996/Final version: 2 October 1996     

A Comparison Between Recipients Receiving Matched Kidney and Those Receiving Mismatched Kidney from the Same Cadaver Donor

The optimal allocation of cadaveric kidneys for transplantation with reference to human leukocyte antigen (HLA) match and sharing these organs to a distant center remains controversial. The current analysis was performed using the United Network for Organ Sharing (UNOS) database for cadaveric kidney transplants (Tx) between 1988 and 1997. The graft survivals of zero-mismatch (matched) kidneys with the mate (mismatched) kidneys were compared. There were 2385 donors and 4770 Tx. Significant differences in recipient demographics between matched and mismatched Tx were: fewer African-American race (AA) in the matched group (9.0% vs. 21.9%), higher number of previous Tx (25.5% vs. 14.8%) and elevated mean cold ischemia time (24.0 vs. 22.2 h). Post-Tx dialysis requirements were similar (22.8% vs. 24.1%, p = 0.62) and matched kidneys had to travel more distance (920 vs. 232 miles). Using a Cox model, the matched group had a decreased relative hazard of graft failure of 23.0% (p = 0.0002) or 35% (p < 0.0001) with and without censoring for death. There was significantly better graft survival in the matched recipients in all pairs except AA (matched) and non-AA (mismatched). For older donors (> or = 50 years, n = 1508), the matched grafts survival was marginally significant (p =0.05). Matched kidneys have improved survival compared with the mismatched kidneys despite the longer distance traveled. The benefit of mismatched transplants was predominantly seen in non-AA.