2459

Magnetic Resonance Imaging of Pulmonary Nodules

Chi Wan Koo¹, Aiming Lu¹, Edwin A Takahashi¹, Jessica Magnuson¹, Peter D Kollasch², Jennifer R Geske³, Julie An⁴, Dennis Wigle⁵, and Tobias Peikert⁶

¹Radiology, Mayo Clinic, Rochester, MN, United States, ²Siemens Medical Solution USA, Inc, Minneapolis, MN, United States, ³Biomedical Statistics and Informatics, Mayo Clinic, Rochester, MN, United States, ⁴Northeast Ohio Medical University, Rootstown, OH, United States, ⁵Thoracic Surgery, Mayo Clinic, Rochester, MN, United States, ⁶Pulmonary and Critical Care Medicine, Mayo Clinic, Rochester, MN, United States

Synopsis

Magnetic resonance imaging had been explored as a potential alternative to computed tomography but the majority of prior MRI nodule studies was performed with 1.5-T scanners and not with the most up to date sequences. Our study demonstrated that biomarkers derived from state of the art 3T MRI sequences can distinguish benign from malignant pulmonary nodules and correlate with morphologic and physiologic values derived from commonly used noninvasive imaging modalities.

Introduction

With the advent of lung cancer screening, a growing number of pulmonary nodules with unknown malignant potential is being detected. Currently, there are no reliable computed tomography (CT) or positron emission tomography (PET)/CT features that distinguish benign from malignant nodules. Therefore, patients often undergo invasive procedures such as biopsies or resections for definitive diagnosis with the disadvantage of post procedural morbidities. Serial follow up CT and/or PET/CT can result in significant additive radiation [1, 2]. Magnetic resonance imaging (MRI) had been explored as an attractive potential alternative to CT due to the lack of ionizing radiation [3]. Recent technologic improvements have addressed past issues rendering MRI to be an attractive viable substitute for CT [4-7]. The majority of prior MRI nodule studies was performed with 1.5-T scanners and not with the most up to date sequences.

Therefore, the aim of our study was to determine whether MR biomarkers derived from up-to-date 3T MR sequences can distinguish malignant from benign pulmonary nodules, differentiate primary from secondary malignancies, and compare MRI-derived measurements with morphologic and physiologic values derived from commonly used noninvasive imaging modalities.

Methods

After informed consent, 45 patients with one or more pulmonary nodules (<3cm) undergoing surgical resection or biopsy were prospectively recruited for 3-T MRI. Respiratory triggered diffusion-weighted imaging (DWI), T2-weighted 2D turbo spin echo with fat saturation, T1-weighted dual-echo 3D VIBE using CAIPIRINHA parallel imaging with and without contrast were performed. Apparent diffusion coefficient (ADC), T1- and T2 signal intensities, enhancement slope, enhancement ratio and enhancement ratio slope derived from region of interest placed on lesions were compared with nodule histology to determine whether these values could distinguish malignant from benign pulmonary nodules and discern primary from secondary malignancies using logistic regression and area under the receiver operating characteristics (ROC) curve. An ADC cutoff was selected to maximize the sum of sensitivity and specificity. These values were compared with PET imaging to examine if they could predict PET positivity, correlate with standard uptake value (SUV) or CT Hounsfield Units (HU). Further associations were tested using chi-square and Fisher’s exact tests for categorical variables, and a t-test to compare T1 visibility with nodule type.

Results

Fifty-one nodules were identified: 8 part-solid and 43 solid. Thirty eight nodules were malignant. There was statistically significant correlation between ADC and malignancy (odds ratio 7.64, p<0.05). An ADC ≥ 1.35 µm2/ms predicted malignancy with 83.3 % specificity (area under curve [AUC] 0.68). While there was no significant correlation between T2 values and malignancy (p=0.18), there was a trend that T1 measurements correlated with malignancy (p=0.06). None of the MRI parameters distinguished primary from metastatic neoplasms. Interestingly, none of the MRI parameters predicted PET positivity but ADC correlated with SUV (p=0.04). Of the 14 malignant nodules that were PET negative, 9 (64%) had an ADC ≥ 1.35 µm2/ms. Both T1 (p < 0.01) and T2 (p = 0.01) values correlated with CT HU measurements and therefore could predict nodule density. All nodules were visible on T2-weighted images while nodule visibility on T1-weighted images depended on CT density (p < 0.01). Nodule type (solid vs. part-solid) did not affect T1 visibility(p=0.32). CT HU (p=0.14) and nodule size (p=0.23) did not affect nodule visibility on DWI. Of the 48 studies with contrast enhancement, 15 demonstrated fat-water swap either before or after contrast enhancement. Lesion type (primary vs. metastatic, p=0.07) but not pathology (p=0.73) might predict this artifact. None of the contrast enhancement parameters correlated with pathology or distinguished primary from metastatic malignancies.

Discussion

Similar to prior studies, our results demonstrated continued capability of ADC to distinguish benign from malignant nodules using updated DWI [7] with a similar ADC cut off for malignancy . We demonstrated that ADC correlates with SUV and more than half of PET negative malignant nodules were positive on DWI signifying that DWI can supplement PET/CT. There is a trend that T1 correlated with malignancy, potentially because T1 correlates with CT density which is increased in high grade adenocarcinomas. Given its high sensitivity for nodules, T2 could be a useful sequence for nodule detection. Although our study did not demonstrate utility of contrast enhancement for nodule characterization, further studies using more advanced dynamic contrast enhancement is warranted

Conclusion

Parameters from up-to-date 3T MRI can distinguish benign from malignant pulmonary nodules and correlate with morphologic and physiologic values derived from commonly used noninvasive imaging modalities. Therefore, MRI biomarkers can serve as a radiation free alternative to commonly used modalities for nodule evaluation.

Acknowledgements

No acknowledgement found.

References

1. Kubo T, Lin PJP, Stiller W, et al. Radiation dose reduction in chest CT: A review. Am J Roentgenol 2008:190;335-343.

2. Fazel R, Krumholz HM, Wang Y, et al. Exposure to low-dose ionizing radiation from medical imaging procedures. N Engl J Med 2009; 361:849-857.

3. Henzler T, Dietrich O, Krissak R, et al. Half-Fourier-Acquisition Single-Shot Turbo Spin-Echo (HASTE) MRI of the lung at 3 Tesla using parallel imaging with 32-receiver channel technology. J Magn Reson Imaging 2009: 30:541–546.

4. Biederer J, Hintze C, Fabel M. MRI of pulmonary nodules: technique and diagnostic value. Cancer Imaging 2008; 8:125–130.

5. Zou Y, Zhang M, Wang Q, Shang D, Wang L, Yu G. Quantitative investigation of solitary pulmonary nodules: dynamic contrast-enhanced MRI and histopathologic analysis. AJR 2008;191:252–9.

6. Koo CW, White DB, Lingineni RK, et al. Magnetic Resonance Imaging of Part-solid Nodules: A Pilot Study. J Thorac Imaging. 2016 Jan;31(1):2-10.

7. Shen G, Ma H, Liu B, Ren P, Kuang A. Diagnostic performance of DWI with multiple parameters for assessment and characterization of pulmonary lesions: A meta-analysis. Am J Roentgenol 2018: 210:1–10.

Proc. Intl. Soc. Mag. Reson. Med. 26 (2018)

2459