Main Session
Sep 29
PQA 04 - Gynecological Cancer, Head and Neck Cancer

2878 - Exploring the Use of Advanced DWI Techniques to Improve Prediction of Early Treatment Response during Definitive Chemoradiation for Cervical Cancer

10:45am - 12:00pm PT
Hall F
Screen: 7
POSTER

Presenter(s)

Vonetta Williams, MD, PhD Headshot
Vonetta Williams, MD, PhD - Memorial Sloan Kettering Cancer Center, New York, NY

V. Y. Yu1, C. Wu1, S. Jogi2, M. A. Kollmeier3, and V. M. Williams3; 1Memorial Sloan Kettering Cancer Center, New York, NY, 2Memorial Sloan Kettering, NY, NY, 3Department of Radiation Oncology, Memorial Sloan Kettering Cancer Center, New York, NY

Purpose/Objective(s): Apparent diffusion coefficient (ADC) from DW MRI have been shown to correlate with response to definitive chemoradiation for cervical cancer. Therefore, we explore the capability of advanced DWI techniques to further improve early treatment response predictions.

Materials/Methods: As part of an IRB-approved study, nine women undergoing chemoradiation for cervical carcinoma obtained MRIs at 7 timepoints (pre-treatment (1), weekly during RT (5) and 3-4 month post treatment (1) ). Two advanced DWI sequences: multi-shot EPI (msEPI) and multi-shot TSE (msTSE), were compared with conventional single-shot EPI (ssEPI). Rigid registration between the T2-weighted (T2w) and b=0 reference scan from each DWI sequence was performed. For all timepoints, a tumor contour based on the T2w scan (TumorT2), and tumor subvolumes for each sequence (SubVolssEPI, SubVolmsEPI, SubVolmsTSE) were delineated. Subvolumes were defined as hypointense regions on the ADC map and hyperintense regions on the high b value scan overlapping TumorT2. The mean ADC values and volume changes of all contours were extracted.

Results: Longitudinal decrease in tumor and subvolume sizes and diffusion signal were observed. Week 2 of RT was chosen for comparison as prior studies identified this timepoint as most indicative of early response. msTSE was noisier but had improved geometric integrity for contouring compared to ssEPI and msEPI, where increased signal pile-up and geometric distortion were seen in the presence of gas. The mean ADC values (average ± std across patient cohort) within TumorT2 at baseline on the msEPI, ssEPI and msTSE scans were (1.20±0.08, 0.94±0.10, 1.12±0.16) 10-3 mm2/s, and the percent increase at week 2 were (20.6±13.2, 21.3±12.9, 18.8±15)%, respectively. This demonstrates that while the absolute ADC values vary between sequences, the change is consistent. The pretreatment volumes of TumorT2, SubVolssEPI, SubVolmsEPI, SubVolmsTSE were (66.3±53.4, 48.2±42.4, 42.6±36.3, 47.7±42.4) cc, and the percent decrease at week 2 of RT were (36.1±9.5, 46.6±20.6, 54.1±22.9, 49.5±19.9)%, respectively. Details are shown in Table 1.

Conclusion: Tumor volume delineation based on DWI sequences may be more sensitive at detecting true changes in residual disease than contours based on T2w anatomical imaging.

Abstract 2878 - Table 1

Percent change at Week 2 of RT

Pt1

Pt2

Pt3

Pt4

Pt5

Pt6

Pt7

Pt8

Pt9

Mean±Std

Tumor ADC change in TumorT2 (%)

ADCmsEPI

16.5

32.4

39.4

25.8

22.9

2.6

1.4

31.3

13.5

20.6 ± 13.2

ADCssEPI

7.3

16.2

44.6

35.2

24.1

9.3

6.7

25.1

23.3

21.3 ± 12.9

ADCmsTSE

9.9

1.7

49.9

23.8

34.1

9.4

9.3

17.3

13.9

18.8 ± 15.0

Percent Volume Decrease (%)

Tumor T2

36.9

38.8

35.6

30.3

54.2

19.2

43.2

31.8

34.9

36.0 ± 9.52

SubVolmsEPI

57.1

31.6

71.5

36.2

67.9

6.1

40.3

60.9

47.8

46.6 ± 20.6

SubVolssEPI

58.5

53.6

77.1

29.5

69.1

30.6

38.8

95.1

34.6

54.1 ± 22.9

SubVolmsTSE

64.4

52.7

83.0

32.9

72.9

34.3

43.1

28.8

32.9

49.4 ± 19.9