Whole-body diffusion-weighted MRI for staging of women with cancer during pregnancy: a pilot study

Abstract

Background: Our aim was to evaluate the feasibility of whole-body diffusion weighted magnetic resonance imaging (WB-DWI) for staging of women with cancer during pregnancy. Material and Methods: Twenty patients diagnosed with cancer during pregnancy underwent WB-DWI in addition to conventional imaging in this prospective single centre study. Reproducibility of WB-DWI between 2 readers was evaluated using Cohen's κ statistics and it's accuracy was compared to clinical imaging for assessing primary tumour site, nodal metastases and visceral metastases. Histopathology after surgery or biopsy was the primary reference standard. Results: Median gestational age at time of WB-DWI scan was 21 weeks (range 7 2/7 to 36 3/7 weeks). Ten patients presented with breast cancer, 3 with lymphoma, 2 with cervical uterine cancer, 1 with ovarian borderline tumour, 2 with colon cancer, 1 with lung cancer and 1 with a conjunctival tumour. Tumour staging by WB-DWI showed very good agreement between both reader with a κ of 0.94 (95% CI 0.81-0.99). Using WB-DWI, reader 1 detected 38 of 41 malignant lesions at a total of 240 anatomical sites and reader 2, 39 lesions compared to 27 for routine clinical imaging. WB-DWI sensitivities were 92.7% (95% CI: 78-98) and 95.1 (95% CI: 82-99) respectively for reader 1 and 2 with specificities of 98.5% (95% CI: 97-99) and 97.5 (95% CI: 95-99). In comparison clinical routine imaging showed sensitivity of 67.5% (95% CI: 57-81) with specificity of 94.2% (95% CI: 97-99). WB-DWI showed highest impact in the detection of nodal and distant metastases with both readers achieving sensitivity of 95% (95% CI: 74-99) with specificity up to 99% (95% CI: 76-99) for nodal staging compared to 50% sensitivity (95% CI: 28-72) with 100% (95% CI: 97-100) specificity for clinical routine imaging. For staging of distant metastases, WB-DWI sensitivities were 66.7% (95% CI: 13-98) and 100% (95% CI: 40-100) respectively for reader 1 and 2 with specificities of 94.1% (95% CI: 69-99) and 100% (95% CI: 40-100). For distant staging; clinical routine imaging showed sensitivity of 33.3% (95% CI: 1.7-87) with specificity of 100% (95% CI: 77-100). Conclusion: WB-DWI is feasible as a single-step non-invasive imaging modality for cancer staging during pregnancy and shows additional value to conventional imaging procedures for detection of distant and nodal metastases.