Yi‐Jen Chen

Identification of vascular hotspots and analysis of micro‐vessel flow velocity waveforms in high‐grade squamous intraepithelial lesions of the cervix

AbstractObjectivesTo assess hotspot micro‐vessel flow velocity waveforms in human papillomavirus (HPV) cervical infections using transvaginal power Doppler ultrasound (TV‐PDU) and explore associations with high‐grade squamous intraepithelial lesions (HSIL, cervical intraepithelial neoplasia [CIN] II and III).MethodsIn all, 62 patients with confirmed HPV‐HSIL (14 CIN II, 48 CIN III) and 65 age‐ and parity‐matched women with neither HPV infection nor CIN were compared. Seven parameters by TV‐PDU were used to assess vascular classification and micro‐vessel flow velocity, including vascular grading (class I, II, III), lowest pulsatility index (PI), resistance index (RI), peak systolic velocity (PS), end‐diastolic velocity (ED), time average maximum velocity (TAMV), and the vascular index (VI = PS/ED).ResultsHSIL was primarily associated with vascular class I (75.8%), followed by class II (14.5%) and class III (9.7%). PI, RI, and VI in HSIL were significantly lower than the control group (P < 0.0001). Mean PI, RI, and VI values decreased progressively from the normal cervix to CIN II–III. At a PI cutoff of 1.03, sensitivity was 88.7%, specificity was 83.8%, and area under the curve (AUC) was 95.0. At an RI cutoff of 0.68, sensitivity was 96.8%, specificity 61.5%, and AUC 84.0. At a VI cutoff of 2.84, sensitivity was 85.5%, specificity 78.5%, and AUC 85.0.ConclusionBased on different patterns of hotspot vascular classification and micro‐vessel flow velocity waveforms, particularly PI between HSIL and the normal cervix, TV‐PDU may offer a potential role for aiding the planning for patients with suspicious HSIL. Further studies are needed to validate the findings.

Comparison of the O-RADS and ADNEX models regarding malignancy rate and validity in evaluating adnexal lesions

This study aimed to compare the ability of the O-RADS and ADNEX models to classify benign or malignant adnexal lesions. This retrospective single-center study included women who underwent surgery for adnexal lesions. Two gynecologists independently categorized the adnexal lesions according to the O-RADS and ADNEX models. Four additional readers were included to validate the new quick-access O-RADS flowchart. Among the 322 patients included in this study, 264 (82.0%) had a benign diagnosis, and 58 (18.0%) had a malignant diagnosis. The malignant rates of O-RADS 2, O-RADS 3, O-RADS 4, and O-RADS 5 were 0%, 3.0%, 37.7%, and 78.9%, respectively. The AUC of the O-RADS in the 322 patients was 0.93. On comparing the O-RADS and ADNEX models in the remaining 281 patients, the AUCs of the O-RADS, ADNEX model with CA125, and ADNEX model without CA125 were 0.92, 0.95, and 0.94, respectively. When setting a uniform cutoff of ≥ 10% (≥ O-RADS 4) to predict malignancy, the O-RADS had higher sensitivity than the ADNEX model (96.6% vs. 91.4%), and relatively similar specificity. In addition, the readers with the quick-access flowchart spent less time categorizing O-RADS than the readers with only the original O-RADS table (mean analysis time: 99 min 15 s vs. 111 min 55 s). The O-RADS classification of the adnexal lesions as benign or malignant was comparable to that of the ADNEX model and had higher sensitivity at the 10% cutoff value. A quick-access O-RADS flowchart was helpful in O-RADS categorization and might shorten the analysis time. • Both O-RADS and ADNEX models had good diagnostic performance in distinguishing adnexal malignancy, and O-RADS had higher sensitivity than ADNEX model in uniform 10% cutoff to predict malignancy. • Quick-access O-RADS flowchart was developed to help review O-RADS classification and might help reduce the analysis time.