1030 - Hypoxia Detection in Prostate Cancer with Oxygen-Enhanced BOLD MRI

M. Swinton¹, D. McHugh², M. Dubec³, C. Moore⁴, A. Sachdeva⁴, D. F. Sanchez⁴, P. Oliveira⁴, C. A. Hart⁵, V. Ramani⁴, M. Lau⁴, J. Oates⁴, N. Clarke⁴, G. Price⁴, A. McWilliam⁶, M. B. van Herk⁷, P. Hoskin⁸, D. L. Buckley⁹, A. Hudson⁴, A. Choudhury⁶, and R. G. Bristow⁶; ¹Christie Hospital, Manchester, United Kingdom, ²Division of Cancer Sciences, The University of Manchester, Manchester, United Kingdom, ³Medical Physics and Engineering, The Christie NHS Foundation Trust, Manchester, United Kingdom, ⁴The Christie NHS Foundation Trust, Manchester, United Kingdom, ⁵University of Manchester, Manchester, United Kingdom, ⁶Division of Cancer Sciences, University of Manchester, Manchester, United Kingdom, ⁷Division of Cancer Sciences, School of Medical Sciences, Faculty of Biology, Medicine and Health, University of Manchester, Manchester, United Kingdom, ⁸Mount Vernon Cancer Centre, Middlesex, United Kingdom, ⁹Christie Medical Physics and Engineering, The Christie NHS Foundation Trust, Manchester, United Kingdom

Purpose/Objective(s): Intra-tumoral hypoxia is associated with a worse prognosis in prostate cancer. We hypothesised that functional MR detection of hypoxia would correlate with in situ biomarkers of hypoxia. If true, hypoxia imaging could identify men that would benefit from treatment intensification. Materials/

Methods: An MR hypoxia detection protocol was run on both a diagnostic 1.5 T MR and 1.5 T MR Linac and included intravoxel incoherent motion (IVIM), oxygen-enhanced (OE) T1 and blood oxygen level-dependent (BOLD) imaging. Imaging was acquired at 2 timepoints prior to radical prostatectomy in the HYPROGEN trial (NCT05702619). Here we report the OE-BOLD data only, quantifying T₂^* values measured while patients inhaled medical air (T₂^*_air) followed by 100% oxygen (T₂^*_O2). Following prostate dissection, histological sections through tumour were stained for GLUT1 expression with a rabbit monoclonal antibody. GLUT1 staining intensity was quantified on HALO software as a tumour GLUT1 H-score (0-300). Median GLUT1 H-score defined the cut-off between low GLUT1 (LG) and high GLUT1 (HG) tumours. Dominant prostatic lesion (DL) and non-malignant prostate (NP) regions were defined on T₂W images and contours applied to T₂^* maps. Median T₂^* values within DL and NP were reported, averaged across both timepoints and compared using a two-sided paired t-test. A two-sided unpaired t-test compared LG and HG regions.
Results: 11/13 patients received 2 MR scans. GLUT1 H-scores available for 9/13 patients and ranged from 95.8 – 194.7 with a median of 144.5. Across all patients, T₂*O₂ values were higher than T₂*air in both NP (p<0.01) and DL (p<0.01). In the HG population, T₂*O₂ was higher than T₂*air in both NP (p=0.02) and DL (p=0.02). In the LG population T₂*O₂ values were higher than T₂*air in the NP (p<0.01) but not in DL (p=0.38). NP values for both T₂*air and T₂*O₂ were higher in the LG population than HG (p<0.01 and p=0.02 respectively). DL values of T₂*air and T₂*O₂ did not differ between the HG and LG populations (p=0.43 and p=0.88 respectively).
Conclusion: Our analysis used the GLUT1 intrinsic biomarker to define intraprostatic hypoxia. We provide evidence T₂* sequences with oxygen challenge can differentiate hypoxic regions as more hypoxic (HG) tumours had lower T₂*air and T₂*O₂ values in NP than less hypoxic (LG) tumours. Further analysis will compare the ability of T₂* sequences to discriminate hypoxic tumours against other MR parameters; OE, IVIM and DCE, using both GLUT1 and the extrinsic hypoxia marker pimonidazole. Table 1. Values for T₂* measured are reported as mean values for each group (standard error)

Group	No. of patients	VOLUME (mm3)		BOLD (T2* air, ms)		BOLD on 100% O2 (T2*O2, ms)		?BOLD with 100% O2 (dT2*, ms)
		NP	DL	NP	DL	NP	DL	NP	DL
All patients	13	40024 (5978)	2591 (611)	55.7 (1.6)	54.8 (1.0)	58.7 (1.6)	56.7 (1.1)	2.98 (0.4)	1.83 (0.5)
HG	5	34413 (8180)	3268 (1548)	52.2 (0.8)	53.9 (2.2)	54.6 (1.4)	56.7 (2.4)	2.49 (0.7)	2.84 (0.8)
LG	4	38490 (6830)	2069 (247)	57.1 (1.0)	56 (0.7)	59.8 (2.0)	57.2 (1.4)	2.77 (0.5)	1.2 (1.2)