Whole-body magnetic resonance imaging with ultra-short echo time sequence versus fluorodeoxyglucose-positron emission tomography/computed tomography in detecting gynecologic metastases

INTRODUCTION

Gynecological cancers (GCas) including cervical, ovarian, and endometrial carcinomas remain a major public health problem that is affecting particularly younger women. There were more than 662,301 new cases of cervical cancer, 420,368 new cases of endometrial cancer, and more than 324,603 new cases of ovarian cancer in 2022 worldwide.^[1] Accurate staging is crucial for predicting patient survival and treatment planning. Positron emission tomography/computed tomography (PET/CT) is the method of choice to assess tumor spread and distant metastases. It has high sensitivity and specificity in detecting distant metastases and has an important role in the surveillance of patients.^[2-4] However, PET/CT involves the injection of a radioactive tracer that exposes patients to ionizing radiation and is therefore not practical for multiple follow-up imaging.

Magnetic resonance imaging (MRI) is evolving into an essential tool in the diagnosis and staging of GCas due to its excellent soft tissue contrast and high spatial resolution.^[5-8] Moreover, diffusion-weighted imaging (DWI) and contrast-enhanced imaging provide additional information about gynecological tumors including endometrial polyps, myometrial invasion, and cervical stromal invasion. The advances in MRI hardware and software have reduced the scan time of whole-body MRI (wbMRI) significantly and made its clinical use possible. Being non-invasive and radiation free in combination with additional functional sequences (DWI and contrast enhanced), wbMRI is becoming a promising tool in staging GCas.^[9]

Gynecological cancers often metastasize to the lungs. Due to magnetic susceptibility effects caused by air tissue interfaces, MRI of lungs has always been challenging. The exceedingly short T2* values of lung parenchyma due to susceptibility effects cause the MR signal to decay rapidly, resulting in low signal-to-noise images. Recently, ultra-short echo time (UTE) approaches have gained popularity in lung imaging.^[10-13] UTE sequences use extremely short echo time in the range of tens of microseconds which allow capturing MRI signal in spite of the short T2* values of lung tissue.^[11] The UTE with a stack-of-spiral acquisition (UTE spiral Volumetric Interpolated Breath-hold Examination [VIBE]) has been applied for both anatomical and functional imaging of lung diseases.^[10,14,15] However, its efficacy as part of a wbMRI protocol has not yet been shown.

Our primary objective was to develop and evaluate a wbMRI protocol with the goal of assessing its diagnostic performance in identifying gynecological metastases relative to PET/CT. As a novel aspect, we implemented the UTE spiral VIBE sequence in a whole-body imaging protocol for the 1^st time. We sought to evaluate the contribution of the UTE spiral VIBE sequence in the detection of metastatic lesions, including lung masses, within the wbMRI protocol, in comparison to combined MRI protocol and PET/CT.

MATERIAL AND METHODS

RESULTS

Five subjects were enrolled in the study. Three of the participants had cervical cancer while two had ovarian cancer.

WbMRI versus PET/CT

A total of 44 findings (including nodules, lesions, and lymph nodes) were detected on FDG-PET images in five patients. Multi-modal imaging of a 61-year-old woman with recurrent ovarian carcinoma is shown for comparison [Figure 1a-d]. Out of 44 findings, 39 lesions were detected on the wbMRI images with a sensitivity of 89%. False negatives in wbMRI included three lung nodules (3/10), one lymph node (1/21), and one bone lesion (1/6). PET/CT and MRI findings are summarized in Table 2.

Close

Figure 1:

A 61-year-old woman with recurrent carcinoma of the ovary. (a) Fluorodeoxyglucose-positron emission tomography/computed tomography (FDG-PET/CT) showing a pulmonary nodule in the right upper lobe (red arrow). (b) Whole-body T1-weighted image (pre-contrast subtracted from post-contrast) (red arrow), (c) pre-contrast ultra-short echo time (UTE) (red arrow), and (d) post-contrast UTE images also demonstrating the pulmonary nodule in the right upper lobe (red arrow).

Export to PPT

Table 2: Findings from PET, PET/CT (considered the gold standard), and MRI, with details for each MRI sequence are presented.

Findings	Count	PET/CT	MRI	PostC T1	DWI/ADC	PreC UTE	PostC UTE
Lung nodule/lesion	10	10	7	7	3	7	7
Lymph node	21	21	20	19	18	18	18
Bone lesion	6	6	5	5	2	4	4
Omental and hypermetabolic implant	2	2	2	2	1	0	0
Soft tissue nodule	3	3	3	3	2	3	3
Uterus mass	1	1	1	1	1	0	1
Breast lesion	1	1	1	1	1	1	1
Total	44	44	39	38	28	33	34

PreC: Pre-contrast, PostC: Post-contrast, UTE: Ultra-short echo time, DWI: Diffusion-weighted imaging, ADC: Apparent diffusion coefficient, MRI: Magnetic resonance imaging, PET/CT: Positron emission tomography/computed tomography, T1: Weighted

There was no statistically significant difference in the performance of PET/CT (100%) and wbMRI (89%) for all findings (P = 0.06).

UTE spiral VIBE performance

Pre-contrast UTE images showed 33 findings (33/44) which was lower than the combined MRI performance (39/44). UTE performance was higher than DWI images (28/44) but lower than post-contrast T1-weighted images (38/44).

Pre-contrast UTE images detected seven out of the 10 lung nodules (7/10), missing the three non-FDG-avid nodules. Multi-modal imaging of a 61-year-old woman with ovarian cancer, depicting a left level IV cervical lymph node is presented [Figure 2a-d]. The accuracy of pre-contrast UTE and post-contrast T1-weighted images in detecting lung nodules was the same, with each detecting seven out of ten (7/10) [Table 2].

Close

Figure 2:

A 61-year-old woman with ovarian cancer. (a) Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography (FDG-PET/CT) demonstrating a left level 4 cervical lymph node metastasis (red arrow). The metastasis is also depicted on (b) post-contrast T1-weighted (red arrow), (c) pre-contrast ultra-short echo time (UTE) (red arrow), and (d) post-contrast UTE images, (red arrow).

Export to PPT

Pre-contrast and post-contrast UTE spiral VIBE images were consistent in detecting all lesions except for a mass in the central uterus, which was visible only in post-contrast UTE and DWI images.

PET displayed all lymph nodes (21/21) whereas wbMRI missed one tracheal lymph node, possibly due to its small size. UTE detected 18 lymph nodes (18/21), missing 2 paratracheal and 1 subcarinal lymph nodes. Multi-modal imaging of a 79-year-old woman with malignant neoplasm of ovaries, demonstrating a mass in the left upper lobe is presented [Figure 3a-c]. Post-contrast T1-weighted imaging displayed 19 lymph nodes (19/21), missing 2 paratracheal lymph nodes.

Close

Figure 3:

A 79-year-old woman with malignant neoplasm of both ovaries. (a) Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography (FDG-PET/CT) demonstrating a mass in the left upper lobe (red arrow). The mass is also depicted on (b) post-contrast T1-weighted (red arrow) and (c) pre-contrast ultra-short echo time (UTE) images (red arrow).

Export to PPT

For bone lesions, PET detected only three out of six (3/6), missing three non-FDG-avid ones. WbMRI identified five (5/6) missing one non-FDG-avid lesion. UTE detected four bone lesions (4/6), also missing two non-FDG-avid ones.

DISCUSSION

Our results showed no statistically significant difference between wbMRI (sensitivity of 89%) and ¹⁸F-FDG-PET/CT in detecting metastatic spread in patients with advanced-stage cervical and ovarian cancers. However, wbMRI missed three lung nodules (3/10), one lymph node (1/21), and one bone lesion (1/6). The lung nodules and the bone lesion were non-FDG-avid. These lesions were approximately 5 mm in length and did not enhance on MRI, suggesting that their small size and lack of contrast may have contributed to their miss on wbMRI. We speculate that higher resolution MRI could potentially detect them although further investigation is needed to confirm this.

This study represents the first exploration of UTE MRI as a whole-body imaging technique. We employed a two-station approach, utilizing a large FOV to image the body from the skull base to the thighs. Each station required approximately 30 s, resulting in a total acquisition time of 1 min. Pre-contrast whole-body UTE imaging detected 33 lesions out of 44 (75% accuracy), demonstrating the potential of this technique for identifying a range of pathologies. Specifically, pre-contrast UTE MRI successfully identified 7 of 10 lung nodules, 18 of 21 lymph nodes, 4 of 6 bone lesions, 3 of 3 soft tissue nodules, and 1 breast lesion whereas it failed to detect 2 implants and 1 mass at the uterus. The UTE sequence demonstrated an 86% accuracy of lung nodule detection compared to CT in a study of 50 nodules by Cha et al.^[12] The low accuracy may be attributed to the small sample size in lung nodules (a total of 10) and the reduced resolution inherent to whole-body UTE applications. Given the relatively short duration of UTE scans, there is potential for optimizing the sequence to achieve improved accuracy even in a whole-body setting.

Interestingly, UTE MRI was able to identify most lesions detected by conventional MRI sequences, including DWI/ ADC and post-contrast T1-weighted images, without the use of a contrast agent. Although the accuracy of pre-contrast UTE MRI (75%) was slightly lower than that of post-contrast T1-weighted images (86%), it was higher than the accuracy (55%) of DWI/ADC sequences. The high performance of UTE MRI can be attributed to its ability to minimize T2* effects which preserves greater proportion of MR signal and its spiral k-space sampling scheme which exhibits low sensitivity to motion artifacts, improving particularly lung imaging.^[16] We conclude that whole-body UTE imaging warrants further investigation to assess its true potential to determine if it can be a viable alternative to contrast-enhanced imaging in cancer staging for patients with renal insufficiency, where the use of gadolinium-based contrast agents is contraindicated.

FDG-PET/CT is a well-established imaging technique in staging GCas.^[7,17] It provides valuable information about cellular metabolism. However, it is invasive and requires injection of radioactive tracers into the patient’s body, which increases radiation exposure and poses a risk for follow-up imaging. In contrast, MRI offers excellent tissue differentiation without radiation risk. WbMRI is a growing technique used to stage patients with various cancers.^[9,18-20] Based on our preliminary results, wbMRI with UTE has the potential to serve patients with GCas who need staging or follow-up imaging. However, the adoption of this technique is hindered by the limited availability of adequate MRI hardware, including whole-body coils, in many imaging centers. As new clinical guidelines are developed, we expect WbMRI with UTE to be introduced into the radiological evaluation of gynecological malignancies.

This study has several limitations. First, our findings are based on a small cohort, primarily due to the inclusion criteria requiring the confirmation of metastatic spread to the lungs through PET/CT for evaluating the UTE sequence. Second, we used 18F-FDG-PET/CT as the gold standard for cancer staging to evaluate the diagnostic performance of wbMRI, due to the absence of an absolute gold standard such as biopsy. Consequently, we were limited to performing sensitivity analysis and could not conduct a specificity analysis since we did not have true negatives. Specificity analysis could have been evaluated if histopathology was used as gold the standard, but this was not possible due to the advanced stage of the disease with multiple metastatic lesions, making biopsy impossible. Finally, patients continued to receive treatment between PET/CT and wbMRI scans, which may have affected some lesions. However, we believe that this did not significantly impact our results, as the patients were in an advanced stage of cancer, with metastases to the lungs and other areas.

CONCLUSION

wbMRI has comparable diagnostic performance to FDGPET/CT for identifying gynecological metastases. Whole-body UTE MRI shows promise as a valuable tool for detecting gynecological metastases without the need for contrast agents. Study findings will need to be confirmed in future research including larger cohorts.