Is the Yedikule-solitary pulmonary nodule malignancy risk score sufficient to predict malignancy? An internal validation study

Abstract

OBJECTIVES We aimed to develop a malignancy risk score model for solitary pulmonary nodules (SPNs) using the demographic, radiological and clinical characteristics of patients in our centre. The model was then internally validated for malignancy risk estimation. METHODS A total of 270 consecutive patients who underwent surgery for SPN between June 2017 and May 2019 were retrospectively analysed. Using the receiver operating characteristic curve analysis, cut-off values were determined for radiological tumour diameter, maximum standardized uptake value and the Brock University probability of malignancy (BU-PM) model. The Yedikule-SPN malignancy risk model was developed using these cut-off values and demographic, radiological and clinical criteria in the first 180 patients (study cohort) and internally validated with the next 90 patients (validation cohort). The Yedikule-SPN model was then compared with the BU-PM model in terms of malignancy prediction. RESULTS Malignancy was reported in 171 patients (63.3%). Maximum standardized uptake value and BU-PM scores were sufficient to predict malignancy (P < 0.001 for both), while the effectiveness of nodule size determined on thoracic computed tomography did not reach statistical significance (P = 0.09). When the Yedikule-SPN model developed with the study cohort was applied to the validation cohort, it significantly predicted malignancy (area under the receiver operating characteristic curve: 0.883, 95% confidence interval: 0.827–0.957, P < 0.001). Comparison of patients in the validation group with Yedikule-SPN scores above (n = 53) and below (n = 37) the cut-off value of 65.75 showed that the malignancy rate was significantly higher among patients with Yedikule-SPN score over 65.75 (86.8% vs 21.6%, P < 0.001, odds ratio = 23.821, 95% confidence interval: 7.805–72.701). When compared with the BU-PM model in all patients, the Yedikule-SPN model tended to be a better predictor of malignancy (P = 0.06). CONCLUSIONS The internally validated Yedikule-SPN model is also a good predictor of the malignancy of SPN(s). Prospective and multicentre external validation studies with large patients’ cohorts are needed.

Factors associated with concurrent malignancy risk among patients with incidental solitary pulmonary nodule: A systematic review taskforce for developing rapid recommendations

“…A total of 180 patients (case/control: 117/63) were in the training set and 90 patients (case/control: 54/36) in the validation set. No significant association was found between it and malignant SPNs (OR: 3.17, 95% CI: 0.89–11.33, p = 0.060) 47 …”

“…A total of 11 studies included smoking history in the analysis and found a significant association between smoking and malignant SPNs (aOR: 2.83, 95% CI: 1.84–4.36, p < 0.001, I 2 = 73.3%; Supplementary Figure S4). 16,20,23,25,26,34,35,42,45,47,49 However, we also noticed one study reporting an opposite association that smoking history was a protective factor for malignant SPNs (OR: 0.28, 95% CI: 0.07–1.13, p = 0.075), 42 which may be attributable to inadequate sample size and patient selection bias (Table 2).…”

“…After the screening of the titles and abstracts, the full text of 192 records were assessed for eligibility. Eventually, 32 studies were included in the systematic review 16,20–50 . The literature selection process was presented in Figure 1.…”

“…Three studies reported a history of extrathoracic malignancy (defined as malignancy diagnosed for more than 5 years). They were published between 1997 and 2021, enrolled a total of 853 patients (case/control: 327/526), and scored between 6.5 and 8 points using the AHRQ checklist 16,45,47 . A pooled analysis of these studies found a significant association between the history of extrathoracic malignancy and malignant SPNs (aOR = 5.66, 95% CI: 2.80–11.46, p < 0.001; I 2 = 0.0%; Supplementary Figure S5).…”

“…The expert panel reached an agreement to combine these four studies with one study analyzing the family history of lung cancer, 47 considering that lung cancer is one significant part of malignant cancer. A pooled analysis found family history of malignancy (including lung cancer) was significantly associated with malignant SPNs (aOR = 3.11, 95% CI: 1.66–5.83, p < 0.001, I 2 = 46.%; Supplementary Figure S8).…”