A Deep-Learning Error Detection System in Radiation Therapy

Delivering radiation therapy based on erroneous or corrupted treatment plan data has previously and unfortunately resulted in severe, sometimes grave patient harm. Aiming to prevent such harm and improve safety in radiation therapy treatment, this work introduces a novel, yet intuitive algorithm for strategically structuring the complex and unstructured data typical of modern treatment plans so their treatment sites may automatically be verified with deep-learning architectures. The proposed algorithm utilizes geometric and dose plan parameters to represent each plan's data as a heat map to feed a deep-learning classifier that will predict the plan's treatment site. Once it is returned by the classifier, a plan's predicted site can be compared to its documented intended site, and a warning raised should the two differ. Using real head-neck, breast, and prostate treatment plan data retrieved at two hospitals in the United States, the algorithm is evaluated by observing the accuracy of convolutional neural networks (ConvNets) in correctly classifying the structured heat map data. Many well-known ConvNet architectures are tested, and ResNet-18 performs the best with a testing accuracy of 97.8% and 0.979 F-1 score. Clearly, the heat maps generated by the proposed algorithm, despite using only a few of the many available plan parameters, retain enough information for correct treatment site classification. The simple construction and ease of interpretation make the heat maps an attractive choice for classification and error detection.

Phrenic Nerve Palsy in Severe Tetanus

Unilateral or bilateral raised hemidiaphragms were observed on chest X-ray in three patients with severe tetanus. Diaphragmatic movement was absent on ultrasonography and fluoroscopy. Nerve conduction study confirmed phrenic nerve palsy. Bilateral involvement caused delayed weaning from the ventilator, whereas unilateral involvement was asymptomatic. There was complete recovery from phrenic nerve palsy in all patients.

A single dose of preoperative gabapentin for pain reduction and requirement of morphine after total mastectomy and axillary dissection: randomized placebo-controlled double-blind trial

BACKGROUND

Gabapentin has been recently found to be useful for reducing acute postoperative pain when administered preoperatively. Although various dose regimens have been tried in different surgical settings, the minimum effective dose is not established.

AIMS

We aimed to evaluate the analgesic efficacy of single low dose gabapentin in patients undergoing total mastectomy and axillary dissection.

SETTINGS AND DESIGN

Prospective randomized placebo-controlled double-blind trial in a tertiary care teaching hospital.

MATERIALS AND METHODS

Fifty women scheduled for total mastectomy and axillary dissection were randomized to receive either gabapentin 600 mg or placebo orally 1 h preoperatively. The intraoperative and postoperative management was standardized. Postoperative pain was assessed at rest and on movement for 12 h using the numerical rating scale (NRS). Morphine was administered if NRS exceeded 30. Primary outcome measure was total morphine consumption.

STATISTICAL ANALYSIS

The morphine consumption was compared using independent t test while pain and sedation scores were analyzed using Mann-Whitney U test.

RESULTS

Forty-six patients completed the trial. The postoperative morphine consumption was significantly less (5.8 +/- 4.2 vs. 11.0 +/- 3.4 mg; P 0.001) and the median [IQR] time to first analgesic was significantly longer (90 [37.5-120] vs. 0 [0-90] min; P 0.001) in the gabapentin group than in the placebo group. The incidence of side effects was similar in the two groups.

CONCLUSIONS

A single low dose of 600 mg gabapentin administered 1 h prior to surgery produced effective and significant postoperative analgesia after total mastectomy and axillary dissection without significant side effects.

Sedation in patients above 60 years of age undergoing urological surgery under spinal anesthesia: comparison of propofol and midazolam infusions

CONTEXT

Propofol and midazolam are commonly used sedatives during regional anesthesia in adults. Smaller doses of these drugs are required in older age due to altered pharmacokinetics and pharmacodynamics.

AIMS

To study the sedation, side-effects and the costs involved with smaller doses of propofol and midazolam in patients aged above 60 years during spinal anesthesia.

SETTINGS AND DESIGN

A randomized single-blind study was conducted in 60 ASA I-II patients aged > or = 60 years undergoing urological surgery under spinal anesthesia.

MATERIALS AND METHODS

Sedation was administered after spinal anesthesia using propofol (bolus 0.4 mg.kg -1; infusion 3 mg/kg/hr) or midazolam (bolus 0.02 mg/kg; infusion 0.06 mg.kg -1.h -1) and titrated to achieve a sedation score of 3 on the modified Observer's Assessment of Alertness/Sedation Scale. Perioperative sedation, hemodynamics and respiratory events were monitored.

STATISTICAL ANALYSIS

The analysis for parametric data was done using Student's unpaired t test and the incidence data using Chi-square test.

RESULTS

The onset (13.0+/-4.2 vs. 18.8+/-4.2 min, P < 0.001) and offset (8.9+/-2.8 vs. 12.5+/-3.5 min, P < 0.001) of sedation were faster and the duration of adequate sedation longer (44.7+/-12.5 vs. 29.8+/-12.9% of total infusion time, P < 0.001) with propofol than midazolam. More patients receiving propofol compared to midazolam had hypotension (16 [50%] vs.4 [14.3%], P= 0.003). Airway obstruction occurred frequently in both the groups. Sedation was significantly more expensive with propofol than midazolam (US$ 9.83 +/- 2.80 vs. US$ 0.33 +/- 0.06, P 0.001).

CONCLUSIONS

Propofol provided better titration and adequacy of sedation than midazolam in patients above 60 years of age, but caused hypotension. Lighter sedation is recommended in this age group.