Of the 27 patients examined,
13 were positive for ureteral localization to VLS,
4 of which had bilateral localization,
thus representing a total of 17 lesions.
Only 1 patient showed intrinsic disease involvement,
while in the remaining 16 localizations only ureterolysis was performed,
due to extrinsic involvement (probably fibroaderential).
Of the 14 patients negative for ureteral localization,
6 patients had localizations at Douglas recess,
7 patients in the ovaries,
3 patients showed intestinal localizations,
2 patients in the utero-sacral ligaments,
1 patient showed bladder localization.
The expert reader obtained sensitivity,
specificity,
VPP and VPN values of 82%,
97%,
0.93 and 0.92 respectively; in particular,
he correctly interpreted the excretory pathways’ involvement in 11/13 patients and 14/17 lesions,
with only one false positive reported.
The inexperienced reader,
on the other hand,
obtained Se,
Sp,
VPN and VPP values of 58.8%,
86.4%,
0.66 and 0.82,
with 9/13 VP for the Patients and 10/17 for the injuries; in addition,
he reported a fair number of FPs,
in particular 5/14.
The distance between the nodules and ureters was variable between 3 mm and 26 mm in the positive patients,
with an average of 11 mm,
and 10 mm and 35 mm in the negative ones,
with an average of 30 mm.
The difference between the two averages was also statistically significant,
with a p-value equal to 0.0005.
From the analysis performed by the "t-student" test and the "ROC curve" we found a statistically significant difference between nodule-ureter distance in positive patients and in the negative ones.
In particular,
the mean distance between the positive patients was 11.2 mm,
while in negative ones was 29.8 mm.
Therefore,
on the basis of the results obtained,
we can state that if the localization of the parametrio,
Douglas or of the peritoneum is placed less than 11 mm from the ureter,
the extrinsic involvement is very probable; while if the distance is greater than 30 mm the ureteral involvement is excluded.
However,
for intermediate values between 11.2 and 29.8,
in the absence of corollary signs,
such as stretching or ureteral dilatation,
it is very difficult to say whether or not the ureter is involved; therefore in these cases MR shows low sensitivity and low VPP.
The analysis of the ROC curve showed that the distance value of the nodule from the ureter with the highest sensitivity and specificity is 21 mm,
with a Se of 82.4% and Sp of 59.5%; however,
interesting values with a good compromise of Se and Sp were found between 21 and 28 mm range (table 1).
The chi-square test showed statistically significant prevalence of positive patients in the <2 cm group compared to the > 3 cm group (p = 0.0019),
but not comparing the former with the 2 and 3 cm group (p = 0.88).