Purpose
To evaluate the efficiency of Heru Prime with its novel artificial intelligence (AI) system-based algorithm in reducing the testing duration compared to the conventional Swedish Interactive Thresholding Algorithm (SITA) strategy.
Methods
48 eyes from 20 patients with visual field defects (7 mild, 6 moderate, and 15 severe per Hodapp-Parrish-Anderson criteria) and 16 normal control subjects (20 eyes) underwent visual field examinations using two protocols: the re:Imagine Strategy (Heru) and the Humphrey Field Analyzer’s (HFA) SITA Standard 24-2 program (Carl Zeiss Meditec). The primary outcome measured was the time taken to complete the test.
Results
The re:Imagine strategy significantly reduced the visual field testing duration compared to the HFA in the control group (3:41 vs. 4:37 minutes, P < 0.001). In the disease group, both methods showed comparable testing times (6:38 vs. 6:36 minutes, P = 0.8949). The table below explores these differences by severity. The Pearson correlation coefficient between all MD values obtained from both devices was 0.9284 (P < 0.001).
Testing duration and mean deviation by group
| Controls (N=20) | Mild (N=7) | Moderate (N=6) | Severe (N=15) | |
|---|---|---|---|---|
| Time (min:s) | ||||
| re:Imagine | 3:41 (0:53) | 5:18 (0:50) | 6:30 (1:01) | 7:19 (0:57) |
| SITA | 4:37 (0:25) | 5:30 (1:07) | 7:13 (0:37) | 6:53 (1:10) |
| P-value (paired t-test) | <0.001 | 0.5532 | 0.2075 | 0.1997 |
| MD (dB) | ||||
| re:Imagine | -2.37 (2.01) | -5.82 (2.04) | -9.33 (3.34) | -17.73 (6.00) |
| SITA | -0.06 (1.37) | -3.65 (1.33) | -8.53 (0.35) | -17.97 (6.00) |
| P-value (paired t-test) | <0.001 | 0.0233 | 0.6087 | 0.8236 |
Values shown as mean (SD). Mild/Moderate/Severe classified per Hodapp-Parrish-Anderson criteria. Disease group total N=28 eyes.
Conclusion
The Heru Prime re:Imagine AI-based algorithm is able to reduce testing duration by predicting patient responses. Heru Prime is faster than SITA Standard in normal patients.
