Skip to main content
Skip to main content
Back to Grand Rounds
Grand RoundsWeekly Evidence Brief

Cardiothoracic Surgery

Edition

30-Second Takeaway

  • Preoperative risk scores can stratify TORS patients for severe 30-day morbidity.
  • ML models show moderate discrimination for in-hospital mortality on ECMO (**AUC 0.759**).

Week ending June 20, 2026

Five recent studies relevant to cardiothoracic surgeons: risk scores, ECMO prognostication, postoperative symptom trajectories, hybrid care trial design, and simulation training evidence

TSOS: preoperative score identifies predictors of severe 30-day morbidity after TORS for OPSCC

ORAL ONCOLOGYJun 18, 2026

In 1,592 TORS patients with OPSCC, 15.7% experienced severe postoperative events within 30 days. Independent preoperative predictors included advanced T stage, BMI <22, smoking, preoperative dysphagia, and hyperlipidemia. A point-based TSOS threshold of ≥24 showed the best discrimination for the composite severe outcome. This is a derivation cohort from TriNetX; external validation is required before clinical deployment.

Machine learning predicts in-hospital mortality for post-cardiotomy ECMO with moderate accuracy

ESC HEART FAILUREJun 15, 2026

Using 5,982 ELSO registry patients, six ML algorithms predicted in-hospital mortality during ECMO support. The boosting model achieved the highest discrimination (AUC = 0.759); other models performed worse. Key predictors included age, on-support lactate, arterial pH, and BMI, but on-support variables reflect trajectory. Authors advise external validation and view models as tools for dynamic risk assessment during ECMO.

Older adults show delayed cough recovery after VATS lung cancer resection compared with dyspnea

BMC GERIATRICSJun 17, 2026

In 292 patients aged 60–80 undergoing VATS, dyspnea largely normalized by 6 months (95.2% mMRC 0–1). Cough remained below preoperative LCQ-MC scores at 6 months and had median recovery time 75 days. VATS lobectomy was associated with slower cough recovery versus sublobectomy (adjusted HR 0.47, 95% CI 0.34–0.66). Use these data to set expectations and plan symptom follow-up in older surgical patients.

References

Numbered in order of appearance. Click any reference to view details.

Additional Reads

Optional additional studies from this edition.

Edition context

Clinical signal

  • Await external validation before using TSOS to alter surgical candidacy or consent.
  • Use ECMO models for dynamic on-support risk assessment, not sole decision-making.
  • Expect slower cough recovery after VATS lobectomy; counsel older patients accordingly.