Unjumbling the TWIST score for testicular torsion: systematic review and meta-analysis

A systematic review and meta-analysis of the studies evaluating the utility of the Testicular Work-up for Ischemia and Suspected Torsion (TWIST) score in establishing or excluding the diagnosis of testicular torsion (TT) is herewith presented in an attempt to quantify the available evidence.

The study protocol was outlined in advance. The review has been conducted in accordance with Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA). The PubMed, PUBMED Central, PMC databases & Scopus followed by Google (Scholar & search engine) were systematically interrogated with the keywords TWIST score, testis and testicular torsion. Fourteen sets of data (n = 1940) from 13 studies were included; data from 7 studies (giving a detailed score-wise break-up) (n = 1285) were dis-integrated and re-integrated to tweak the cut-offs for low and high risk.

For every 4 patients presenting to the Emergency Department (ED) with acute scrotum, one patient will eventually be diagnosed with TT. The mean TWIST score was higher in patients with testicular torsion (5.13 ± 1.53 vs 1.50 ± 1.40 for those without TT). TWIST score can be used to predict testicular torsion at cut-off of 5 with a sensitivity, specificity, PPV, NPV, and accuracy of 0.71 (0.66, 0.75; 95%CI), 0.97 (0.97, 0.98; 95%CI), 90.2%, 91.0%, and 90.9% respectively. While the slider for cut-off was shifted from 4 to 7, there was a rise in specificity and PPV of the test with a corresponding decline in sensitivity, NPV, and accuracy. The sensitivity witnessed a sharp decline from 0.86 (0.81-0.90; 95%CI) @ cut-off 4 to 0.18 (0.14-0.23; 95%CI) @ cut-off 7. The area under the SROC curve for cut-off 5 was more than that for cut-offs 4, 6 & 7. TWIST cut-off of 2 may be used to predict the absence of testicular torsion with a sensitivity, specificity, PPV, NPV, and accuracy of 0.76 (0.74, 0.78; 95%CI), 0.95 (0.93, 0.97; 95%CI), 97.9%, 56.5%, and 80.7%, respectively. While the cut-off is lowered from 3 to 0, there is a corresponding rise in the specificity and PPV, while the sensitivity, NPV, and accuracy are compromised. The sensitivity witnesses a sharp decline from 91 to 35%. The area under the SROC curve for cut-off 2 was more than that for cut-off @ 0, 1 or 3. The sum of sensitivity and specificity of TWIST scoring system to ascertain the diagnosis of TT is more than 1.5 for cut-off values 4 & 5 only. The sum of sensitivity and specificity of TWIST scoring system to confirm the absence of TT is more than 1.5 for cut-off values 3 & 2 only.

TWIST is a relatively simple, flexible, and objective tool which may be swiftly administered even by the para-medical personnel in the ED. The overlapping clinical presentation of diseases originating from the same organ may prevent TWIST from absolutely establishing or refuting the diagnosis of TT in all the patients with acute scrotum. The proposed cut-offs are a trade-off between sensitivity and specificity. Yet, the TWIST scoring system is immensely helpful in the clinical decision-making process and saves time-lag associated with investigations in a significant majority of patients.