A machine learning case study to predict rare clinical event of interest: imbalanced data, interpretability, and practical considerations.

Accurate prediction of a rare and clinically important event following study treatment has been crucial in drug development. For instance, the rarity of an adverse event is often commensurate with the seriousness of medical consequences, and delayed detection of the rare adverse event can pose significant or even life-threatening health risks to patients. In this machine learning case study, we demonstrate with an example originated from a real clinical trial setting how to define and solve the rare clinical event prediction problem using machine learning in pharmaceutical industry. The unique contributions of this work include the proposal of a six-step investigation framework that facilitates the communication with non-technical stakeholders and the interpretation of the model performance in terms of practical consequences in the context of patient screenings for conducting a future clinical trial. In terms of machine learning methodology, for data splitting into the training and test sets, we adapt the rare-event stratified split approach (from scikit-learn) to further account for group splitting for multiple records of a patient simultaneously. To handle imbalanced data due to rare events in model training, the cost-sensitive learning approach is employed to give more weights to the minor class and the metrics precision together with recall are used to capture prediction performance instead of the raw accuracy rate. Finally, we demonstrate how to apply the state-of-the-art SHAP values to identify important risk factors to improve model interpretability.

Structured connectivity in the output of the cerebellar cortex.

The spatial organization of a neuronal circuit is critically important for its function since the location of neurons is often associated with function. In the cerebellum, the major output of the cerebellar cortex are synapses made from Purkinje cells onto neurons in the cerebellar nuclei, yet little has been known about the spatial organization of these synapses. We explored this question using whole-cell electrophysiology and optogenetics in acute sagittal cerebellar slices to produce spatial connectivity maps of cerebellar cortical output in mice. We observed non-random connectivity where Purkinje cell inputs clustered in cerebellar transverse zones: while many nuclear neurons received inputs from a single zone, several multi-zonal connectivity motifs were also observed. Single neurons receiving input from all four zones were overrepresented in our data. These findings reveal that the output of the cerebellar cortex is spatially structured and represents a locus for multimodal integration in the cerebellum.

Safety and efficacy of dexamethasone intravitreal implant for treatment of macular edema secondary to retinal vein occlusionin Chinese patients: randomized, sham-controlled, multicenter study / 中华眼底病杂志

Objective To evaluate the safety and efficacy ofdexamethasone intravitreal implant 0.7 mg (DEX) for treatment of macular edema associated with retinal vein occlusion (RVO).Methods This study was a six-month,randomized,double-masked,sham-controlled,multicenter,phase 3 clinical trial with a 2-month open-label study extension.Patients with branch or central RVO received DEX (n=129) or sham procedure (n=130) in the study eye at baseline;all patients who met re-treatment criteria received DEX at month 6.Efficacy measures included Early Treatment Diabetic Retinopathy Study (ETDRS),best-corrected visual acuity (BCVA),and central retinal thickness (CRT) on optical coherence tomography.Results Time to ＞ 15-letter BCVA improvement from baseline during the first 6 months (primary endpoint) was earlier with DEX than sham (P＜ 0.001).At month 2 (peak effect),the percentage of patients with ≥ 15-letter BCVA improvement from baseline was DEX:34.9％,sham:11.5％;mean BCVA change from baseline was DEX:10.6± 10.4 letters,sham:1.7 ± 12.3 letters;and mean CRT change from baseline was DEX:-407 ± 212 μm,sham:-62 ± 224 μm (all P＜0.001).Outcomes were better with DEX than sham in both branch and central RVO.The most common treatment-emergent adverse event was in-creased intraocular pressure (IOP).Increase sin IOP generally were controlled with topical medication.Mean IOP normalized by month 4,and no patient required incisional glaucoma surgery.Conclusions DEX had a favorable safety profile and provided clinically significant benefit in a Chinese patient population with RVO.Visual and anatomic outcomes were improved with DEX relative to sham for 3-4 months after a single implant.