A clinical practical model for preoperative prediction of visual outcome for pituitary adenoma patients in a retrospective and prospective study

Front Endocrinol (Lausanne). 2024 Dec 13:15:1479442. doi: 10.3389/fendo.2024.1479442. eCollection 2024.

Abstract

Objective: Preoperative prediction of visual recovery after pituitary adenoma resection surgery remains challenging. This study aimed to investigate the value of clinical and radiological features in preoperatively predicting visual outcomes after surgery.

Methods: Patients undergoing endoscopic transsphenoidal surgery (ETS) for pituitary adenoma were included in this retrospective and prospective study. Preoperative MRI, visual acuity, visual field, and postoperative visual recovery data were collected. Logistic regression analysis was used to assess the importance of clinical and MRI features, and a prediction model was developed.

Results: The cohort included 198 patients (150 retrospective, 48 prospective). In the retrospective data, visual recovery was observed in 111 patients (74.0%), while non-recovery was observed in 39 patients (26.0%). In the prospective data, visual recovery was observed in 27 patients (56.25%) and non-recovery in 21 patients (43.75%). Blindness, headache, adenoma area, and adenoma upward growth distance were negatively correlated with visual recovery (p < 0.05), while the pituitary gland area was positively correlated (p = 0.001). Logistic regression selected three clinical features: blindness, headache, and visual impairment course. Two additional imaging features, pituitary gland maximum area, and adenoma maximum area, were incorporated into the prediction model. The area under the curve of the prediction model was 0.944 in the retrospective cohort and 0.857 in the prospective cohort. Accuracy was 88% and 81.25%, respectively.

Conclusion: This study successfully developed a clinical practical model combining clinical and radiological features to preoperatively predict visual recovery for patients with pituitary adenoma. The model has the potential to provide personalized counseling for individual patients.

Keywords: clinical practical model; graphic segmentation; machine learning (ML); pituitary adenoma; visual outcome.

MeSH terms

  • Adenoma* / diagnostic imaging
  • Adenoma* / pathology
  • Adenoma* / surgery
  • Adult
  • Aged
  • Female
  • Humans
  • Magnetic Resonance Imaging*
  • Male
  • Middle Aged
  • Pituitary Neoplasms* / diagnostic imaging
  • Pituitary Neoplasms* / pathology
  • Pituitary Neoplasms* / surgery
  • Prognosis
  • Prospective Studies
  • Recovery of Function
  • Retrospective Studies
  • Treatment Outcome
  • Vision Disorders / etiology
  • Vision Disorders / physiopathology
  • Visual Acuity / physiology
  • Young Adult

Grants and funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. This work was sponsored by Joint Funds for Innovation of Science and Technology, Fujian Province (No. 2021Y9089), University-Industry Research Joint Innovation Project of Science, Technology, Fujian Province (No. 2023Y4018), Fujian Province Finance Project (No. BPB-2022YXR) and KCXFZ20201221173202007.