Opportunities and challenges in interpretable deep learning for drug sensitivity prediction of cancer cells.

In precision oncology, therapy stratification is done based on the patients' tumor molecular profile. Modeling and prediction of the drug response for a given tumor molecular type will further improve therapeutic decision-making for cancer patients. Indeed, deep learning methods hold great potential for drug sensitivity prediction, but a major problem is that these models are black box algorithms and do not clarify the mechanisms of action. This puts a limitation on their clinical implementation. To address this concern, many recent studies attempt to overcome these issues by developing interpretable deep learning methods that facilitate the understanding of the logic behind the drug response prediction. In this review, we discuss strengths and limitations of recent approaches, and suggest future directions that could guide further improvement of interpretable deep learning in drug sensitivity prediction in cancer research.

Optic disc doubling.

Many developmental defects can simulate the optic disc (optic disc pseudodoubling). True optic disc duplication with 2 independent retinal vasculatures is rare. Associated optic nerve duplication is rarer still and seen in lower vertebrates. We report a rare case of unilateral double optic discs with a single optic nerve in the left eye of a 41-year-old man, otherwise asymptomatic in the affected eye.

Optic disc tuberculoma in a patient with miliary tuberculosis.

A 27-year-old female, on anti-tubercular therapy for miliary tuberculosis for 1 week, presented with gradual diminution of right eye vision for 4 months. Right eye visual acuity was counting fingers at 15 cm. Right fundus showed a yellowish-white vascularized lesion, 4 disc diameters in size, overlying the optic disc and associated neurosensory macular detachment. The left eye was normal. A Mantoux test was negative. The authors diagnosed presumed optic disc tuberculoma in the right eye. Oral prednisolone in tapering doses was added to the anti-tubercular therapy. The optic disc lesion regressed with residual scarring and vision became 20/30 at 7 months.

Is inclusion of Sabouraud dextrose agar essential for the laboratory diagnosis of fungal keratitis?

PURPOSE: To determine whether the inclusion of Sabouraud dextrose agar (SDA) is essential in the diagnosis of fungal keratitis. MATERIALS AND METHODS: Corneal scrapings of 141 patients with microbial keratitis were smeared and cultured. Sheep blood agar (BA), chocolate agar (CA), SDA, non-nutrient agar (NNA) with Escherichia coli overlay, and brain heart infusion broth (BHI) were evaluated for time taken for growth and cost. The media were also evaluated experimentally for rate of growth and time taken for identification. RESULTS: Twenty-six of 39 patients positive for fungus in corneal scrapings by microscopy were culture-positive. Fungus grew on BA in 22/39, on CA in 18/39, on SDA in 17/39, on NNA in 17/39, and on BHI in 13/39 cases. Growth on SDA was higher in ulcers with larger infiltrate (6/18 versus 9/13, P = 0.04). Estimated saving with inclusion of only BA/CA was Rs. 600 per patient. Performance of all media was similar in in vitro experiment although the characteristic spores and color were seen earlier on SDA. CONCLUSION: Fungal keratitis can be reliably confirmed on BA or CA, which support growth of both bacteria and fungus.