Comparison of Riete and Some Scores for Predicting Cancer in Unprovoked Thrombosis

Background:

Venous thromboembolism (VTE) can often be the first sign of cancer. The Riete score and the SOME score are clinical prediction tools used to assess the risk of cancer in patients with unprovoked VTE. Although these scores have been studied individually for cancer prediction, no study has compared both scores directly. The aim of this study is to validate the Riete and Some scores and compare their performance in predicting cancer in patients presenting with unprovoked VTE.

Methods:

We conducted a retrospective study of patients diagnosed with unprovoked VTE, defined according to ISTH, from January 2016 to December 2022. VTE includes all pulmonary embolisms and deep vein thrombosis of the lower extremities. Patients with provoked VTE, cancer at diagnosis, and those who were lost to follow-up were excluded. Data were collected through manual chart review. The Riete score items included sex, age, chronic lung disease, anemia, platelet count, prior VTE, and recent surgery. The SOME score items included age , current smoking, and previous provoked VTE. The Riete score was considered high if greater than 2, and the SOME score was considered high if greater than 1. Continuous variables were expressed as medians with interquartile ranges (IQR), and categorical variables were expressed as frequencies and percentages. The performance of the Riete score and SOME score was evaluated by calculating sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for dichotomized variables, while discriminative ability was assessed using the area under the receiver operating characteristic (ROC) curve (AUC) for the continuous scores. The comparison between the sensitivity and specificity of the two scores was performed using McNemar's test, while the DeLong test was used to compare the AUCs of the two scores. Statistical analyses were conducted using SPSS version 29.0.2, with a significance level set at p < 0.05. The study was approved by the Institutional Review Board, and informed consent was waived.

Result

Of the initially identified patients, 1230 were excluded due to having provoked VTE, and an additional 23 were excluded due to loss of follow-up. Consequently, 427 patients with unprovoked VTE were included in the study. The median age of was 60 years (IQR 22 ), with 40.9% being male. Majority (73%) had at least one comorbidity, with hypertension being the most common (41%). Nineteen patients (4.4%) were diagnosed with cancer within 12 months, with the most common types being colorectal cancer (31.6%) followed by lung cancer (15.8%). Majority of patients diagnosed with cancer had localized disease (58 %).

Among the study population, 33.7% were categorized as high risk according to the Riete score, while 46.4% were categorized as high risk according to the SOME score. The median Riete score was 2.0 (IQR 2) with a sensitivity of 42.1%, specificity of 66.7%, PPV of 5.6%, NPV of 96.1%, and an AUC of 0.60. The median SOME score was 0.0 (IQR 1) with a sensitivity of 63.2%, specificity of 54.4%, PPV of 6.1%, NPV of 96.9%, and an AUC of 0.60. No significant difference in sensitivity (p=0.34) or PPV (p=0.34) between the scores, but significant differences were found in specificity (p=0.03) and NPV (p=0.03). The DeLong test for comparing the AUCs yielded a p-value of 0.55, indicating no significant difference between the two scores.

Conclusions

The Riete and SOME scores show modest and similar ability to predict cancer in patients with unprovoked VTE. While the Riete score offers higher specificity, the SOME score provides better sensitivity. To our knowledge, this is the first study to directly compare and validate both scoring systems in exclusively unprovoked VTE patients. Future research with larger sample sizes and prospective validation studies is needed to confirm the predictive ability of these tools. Implementing robust and validated scoring systems could improve early cancer detection and patient outcomes by aiding clinicians in risk stratification, identifying patients who may benefit from extensive screening, and avoiding unnecessary screening in low-risk patients, ultimately leading to more effective cancer surveillance.

Kearon, C et al. JTH vol. 14,7 (2016)

Jara-Palomares, Luis et al. Chest vol. 151,3 (2017)

Ihaddadene, Ryma et al. Blood vol. 127,16 (2016)