Background and Aims Determining clinical probability of pulmonary embolism (PE) with Wells scoring system is the first step towards diagnosis of PE. Definitive diagnosis of PE is confirmed by computed tomography pulmonary angiography (CTPA). Methods This was a prospective study on 80 patients referred to the Institute for Pulmonary Diseases of Vojvodina with suspected PE between April 2010 and August 2012. Clinical probability of PE was determined according to the Wells and modified Wells scoring system. CTPA was performed in 60 patients. The degree of pulmonary vascular obstruction was quantified by the Qanadli index. Results Low clinical probability of PE was present in one patient (1.6%), moderate in 43 (71.6%) and high in 16 (26.6%) patients. PE was confirmed in 50 (83.3%) patients. There were 21 patients (42%) whose Quanadli index was <25%, 18 (36%) between 25%–50%, while Quanadli index was ≥50 in 11 patients (22%). When compared to CTPA findings, modified Wells scoring system showed 90% sensitivity 95% confidence interval (CI) 78.2%–96.6%, and 20% specificity (95% CI 3.11%–55.6%), positive predictive value (PPV) 84.9% (95% CI 72.4%–93.2%) and negative predictive value (NPV) 28.6% (95% CI 4.5%–70.7%). There was weak positive correlation between Wells score and Quanadli index (r = 0.14; P = 0.29), without statistical significance. Wells score was significantly higher in haemodynamically unstable than in haemodynamically stable patients (6.8 vs 5.6, P = 0.014). There was no statistically significant difference between the values of Quanadli index in these two groups (31.33% vs 26.64%, P = 0.062). Conclusion Modified Wells criteria have high sensitivity but low specificity in PE diagnostics. The Wells score does not correlate well with the Quanadli index.