Pad´e quantification of malignant and benign ovarian MRS data - Signal Processing in MRS with Biomedical Applications - page 279

39 cancerous and 53 to 57 benign lesions 2 of various histopathology, as well
as one borderline cancerous adnexal mass that have been examined by means
of in vivo MRS.
The two most recent investigations [334, 339] of benign and malignant ovar
ian lesions were performed using a 3T MR scanner, whereas the other studies
[335]-[338] utilized 1.5T scanners. A limited number of peaks were visible and
these were assigned as follows: lipid at 1.3 ppm and an inverted lactate dou
blet also at around 1.3 ppm, creatine at 3.0 ppm, choline at 3.2 ppm and lipid
at 5.2 ppm. A prominent but unassigned resonance at 2.07 ppm was reported
by Stanwell et al. [339]. The metabolite concentrations were estimated quali
tatively (present or absent) or semiquantitatively (strongly present, present,
possibly present or absent) 3 . Ratios were reported in some of the studies.
Choline at 3.2 ppm was detected in 27 of the 32 malignant ovarian lesions
(84.4%) versus 14 to 18 of 30 to 34 benign ovarian lesions (41.2% 60%).
Stanwell et al. [339] reported that choline to creatine ratios above 3 were
found for all seven malignant ovarian lesions, whereas this ratio was below
1.5 for six of the seven benign lesions (in the 7th benign lesion, a serous
cystadenofibroma, the Cho/Cr ratio was 3.13).
In eleven of the eighteen cancerous ovarian lesions from Refs. [335, 337, 338],
lipid at 1.3 ppm was reported (61.1%) and in twenty of thirtythree non
malignant lesions of the ovary (60.6%). One of the studies [335] also noted
the presence of a lipid peak at 5.2 ppm only among patients with certain non
malignant processes affecting the ovary. The peak at 5.2 ppm was seen in 1
of 4 patients with endometriosis, 2 of 11 patients with teratomas, and in the
single patient with an ectopic pregnancy, but in none of the 6 patients with
benign epithelial tumors nor in the single patient with salpingitis.
Data concerning the inverted lactate doublet at 1.3 ppm were provided in
three of the studies [337]-[339], altogether this doublet was observed in 11
of 18 (61.1%) and in 7 of 27 (25.9%) of the malignant and benign ovarian
lesions, respectively. In Refs. [338, 339] the amplitude of the lactate doublet
was observed to be larger in the malignant compared to the benign lesions.
The unassigned resonance at 2.07 ppm was detected in two of seven (28.6%)
ovarian cancers, and in four of the seven (57.1%) benign lesions (teratomas
and serous cystadenomas) [339].
Based upon this summary of the reported findings to date from in vivo
MRS, it can be concluded that none of the qualitatively or semiquantitatively
assessed metabolites provided su cient distinction between malignant and
benign ovarian lesions. Metabolite ratios (choline to creatine) were reported
in a total of only fourteen lesions [339] and yet there was an overlap between
2 In Ref. [334] the total number of benign gynecological lesions is given, and the various
types described. From this description it can be deduced that at least two and at most six
of these lesions were ovarian.
3 Not all the studies reported on all of the mentioned metabolites. For example, choline was
not described in any way with respect to the thirteen patients examined in Ref. [335].