Trends and outcomes of optical coherence tomography use: 877 patients single-center experience

Development of machine learning models for fractional flow reserve prediction in angiographically intermediate coronary lesions

BACKGROUND

Fractional flow reserve (FFR) represents the gold standard in guiding the decision to proceed or not with coronary revascularization of angiographically intermediate coronary lesion (AICL). Optical coherence tomography (OCT) allows to carefully characterize coronary plaque morphology and lumen dimensions.

OBJECTIVES

We sought to develop machine learning (ML) models based on clinical, angiographic and OCT variables for predicting FFR.

METHODS

Data from a multicenter, international, pooled analysis of individual patient's level data from published studies assessing FFR and OCT on the same target AICL were collected through a dedicated database to train (n = 351) and validate (n = 151) six two-class supervised ML models employing 25 clinical, angiographic and OCT variables.

RESULTS

A total of 502 coronary lesions in 489 patients were included. The AUC of the six ML models ranged from 0.71 to 0.78, whereas the measured F1 score was from 0.70 to 0.75. The ML algorithms showed moderate sensitivity (range: 0.68-0.77) and specificity (range: 0.59-0.69) in detecting patients with a positive or negative FFR. In the sensitivity analysis, using 0.75 as FFR cut-off, we found a higher AUC (0.78-0.86) and a similar F1 score (range: 0.63-0.76). Specifically, the six ML models showed a higher specificity (0.71-0.84), with a similar sensitivity (0.58-0.80) with respect to 0.80 cut-off.

CONCLUSIONS

ML algorithms derived from clinical, angiographic, and OCT parameters can identify patients with a positive or negative FFR.

Optical Coherence Tomography Measures Predicting Fractional Flow Reserve: The OMEF Study

Background

Optical coherence tomography (OCT) allows to carefully characterize coronary plaque morphology and lumen dimensions. We sought to evaluate the value of OCT in predicting fractional flow reserve (FFR).

Methods

We performed a multicenter, international, pooled analysis of individual patient-level data from published studies assessing FFR and OCT on the same vessel. Data from stable or unstable patients who underwent both FFR and OCT of the same coronary artery were collected through a dedicated database. Predefined OCT parameters were minimum lumen area (MLA), percentage area stenosis (%AS), and presence of thrombus or plaque rupture. Primary end point was FFR ≤0.80. Secondary outcome was the incidence of major adverse cardiac events in patients not undergoing revascularization based on negative FFR (>0.80).

Results

A total of 502 coronary lesions in 489 patients were included. A significant correlation was observed between OCT-MLA and FFR values (R = 0.525; P < .001), and between OCT-%AS and FFR values (R = -0.482; P < .001). In Receiver operating characteristic analysis, MLA <2.0 mm2 showed a good discriminative power to predict an FFR ≤0.80 (AUC, 0.80), whereas %AS >73% showed a moderate discriminative power (AUC, 0.73). When considering proximal coronary segments, the best OCT cutoff values predicting an FFR ≤0.80 were MLA <3.1 mm2 (AUC, 0.82), and %AS >61% (AUC, 0.84). In patients with a negative FFR not revascularized, the combination of lower MLA and higher %AS had a trend toward worse outcome (which was statistically significant in the analysis restricted to proximal vessels).

Conclusions

OCT lumen measures (MLA, %AS) may predict FFR, and different cutoffs are needed for proximal vessels.