Monoexponential and Biexponential Models Based Diffusion Weighted Imaging: A Comparative Study in the Differential Diagnosis of Benign and Malignant Solitary Pulmonary Tumors / 中国医学影像学杂志

ABSTRACT

Purpose To explore the value of diffusion weighted imaging monoexponential and biexponential modelsin differentiating benign and malignant solitary pulmonary tumors.Materials and Methods Seventy-seven patients with pathologically confirmed solitary pulmonary tumors were enrolled in this study,a total of 96 lesions were studied,48 benign ones and 48 lung cancers (LC),including 19 cases of squamous cell carcinomas (SCC) and 28 cases of adenocarcinomas (Adeno-Ca).Apparent diffusion coefficient (ADC),slowapparentdiffusion coefficient (Dslow),fast apparentdiffusion coefficient (Dfast) and fast diffusion ratio (f) were calculated with monoexponential and biexponential models for all the pulmonary tumors.Results The ADC and Dslow values of benign masses [ADC (1.37 ± 0.48)× 10-3 mm2/s;Dslow (1.34 ± 0.45)× 10-3 mm2/s] were both significantly higher than those of Adeno-Ca [ADC (1.19±0.23)×10-3 mm2/s;Dslow (1.12±0.35)×10-3 mm2/s] and SCC [ADC (1.13±0.28)×10-3 mm2/s;Dslow (1.02 ± 0.32)× 10-3 mm2/s].There was no significant difference in ADC and Dslow values between Adeno-Ca and SCC (P＞0.05).The Dfast and f values of malignant masses were higher than benign ones,but the difference was not statistically significant (P＞0.05).Dslow were both significantly higher in accuracy (72.2％) and sensitivity (91.3％) than other imaging indicators (accuracy55.5％-68.0％;sensitivity41.3％-78.3％;all P＜0.01) in discriminating LC from benign masses showed by receiver operating characteristic curve analysis.Conclusion DWI biexponential model-derived Dslow shows greater accuracy than ADC in differentiating benign and malignant solitary pulmonary lesions.