Probing Tumor Oxygenation Response to Hypoxic Gas Breathing
Donghan Mo Yang1, Tatsuya J Arai1, James Campbell1, and Ralph P Mason1

1Radiology, The University of Texas Southwestern Medical Center, Dallas, TX, United States


Tumor oxygenation response to hypoxic gas breathing is studied in rat 13762NF breast tumors using BOLD and TOLD MRI, with hyperoxic gas breathing as reference. Time course of R2* and R1-weighted signal is analyzed for 100%, 16%, and 14% O2 breathing challenges, respectively. Evidence for decreased blood oxygen saturation (sO2) and decreased tissue oxygen partial pressure (pO2) is observed, revealing different patterns depending on the tumor size.


Hypoxia in breast tumors is the target of a novel therapeutic strategy deploying bioreductively activatable prodrug conjugates. A prerequisite for developing this therapy is to evaluate the hypoxia profile in various breast tumor models in rodents and, if necessary, to aggravate hypoxia via hypoxic gas breathing. Compared with the extensively studied hyperoxic gas breathing1, hypoxic gas breathing provokes different physiological changes, and hence impacts MRI parameters through different patterns. Herein, we apply oxygenation-sensitive MRI methods to investigate the response to hypoxic gas breathing in rat 13762NF tumors.


Rat 13762NF tumors were implanted subcutaneously into the right thigh of Fischer rats (n = 8) and grown to 0.1–3.5 cm3 in volume. Interleaved R2* (BOLD) mapping and short-TR, R1-weighted, gradient-echo (TOLD) imaging (scan time 41.4 and 16 s, respectively) at 4.7 T was performed across air to 16% O2 and air to 100% O2 breathing challenges (n = 4 in each challenge). In some experiments, R1 maps (8 min with 5 TI’s) were collected before and after the interleaved BOLD and TOLD time course. All images were acquired from a 2-mm slice crossing the center of the tumor with in-plane spatial resolution 0.202–0.312 mm2/pixel (128 × 128), depending on the tumor size. For one additional rat, hypoxic gas breathing challenge was extended to 14% O2 following the regular 16% O2 step. MRI parameters, i.e., R2*, R1-weighted (TOLD) signal intensity (SI), and R1, were analyzed voxel by voxel in tumor and thigh muscle. R2* baseline (air breathing) voxel-wise temporal variation was used as the criterion to distinguish the responsive vs. less responsive areas in the tumor for each gas breathing challenge. Raw TOLD SI was modified, voxel-by-voxel at each dynamic time point, to remove the R2*-weighted signal decay according to the following equation: SI(modified) = SI(raw) / exp(-R2*∙TE), where TE = 5 ms.


Rat 13762NF tumors are found to be highly heterogeneous and develop mucinous necrotic cores when grown larger than 0.5 cm3. Intratumoral responsive area, derived from R2* dynamics, is typically located in the tumor periphery (Fig. 1a), corresponding to viable tumor tissue with accessible vasculature (Fig. 1b). BOLD and TOLD responses to gas breathing challenges in the responsive area of tumor are invariably larger than those in muscle (Fig. 2). Progressive hypoxic gas breathing (air to 16% and then 14% O2) led to a linear correlation between ∆R2* and TOLD SI in tumor, but minimal change was observed in muscle (Fig. 3). Both regional decreased and increased TOLD SI during hypoxic gas breathing was observed in small (~0.1 cm3) and large (>0.4 cm3) tumors, respectively (Fig. 4). Similar magnitude of ∆R2*, yet in different directions, were observed during hypoxic and hyperoxic gas breathing (Fig. 5a).


Tissue R2* is positively correlated with the fraction of deoxyhemoglobin (deoxyHb) in the total hemoglobin in blood. TOLD SI, essentially R1-weighted signal, depends on the combined contribution from blood flow, tissue pO2, and blood deoxyHb fraction1, where molecular O2 and deoxyHb both act as R1 contrast agent. The effect of hyperoxic gas breathing, which has been extensively studied in various types of tumors, is generally decrease in R2* (due to deoxyHb decrease) and increase in TOLD signal (increase in R1 due to pO2 increase), the extent of which further depends on the native oxygenation status and vasculature function of the tissue (or tumor)2-4. Consistent results for 100% O2 breathing are shown here in 13762NF tumors. Hypoxic gas breathing enhances deoxyHb fraction in blood, demonstrated herein by the increased R2*. TOLD response differs for different tumor sizes. For large tumors, an increase in TOLD SI might be dominated by increased deoxyHb which overcomes the opposite impact of decreased pO2 (if any). For small tumors, the decrease in TOLD SI in the presence of evident deoxyHb increase (revealed by the increased R2*) might reflect a large decrease in tumor pO2. Interestingly, during the progressive 16% and 14% O2 breathing challenge, the two independently measured parameters, R2* and TOLD SI, correlate linearly (Fig. 3), implying similar underlying contributor(s). A schematic illustration of possible BOLD and TOLD response pathways is shown in Fig 5b. Future studies will address the role of tumor size and vascular perfusion in the observed BOLD and TOLD responses.


This work explores the less studied effect of hypoxic gas breathing on oxygenation-sensitive MRI parameters in rat 13762NF breast tumors. Combined dynamic BOLD and TOLD measurement proves to be useful for probing changes in tumor oxygenation induced by hypoxic gas breathing.


Supported in part by CPRIT RP-140399, NIH P41-EB015908 and P30-CA142543.


1. Remmele, S.; Mason, R. P.; O’Connor, J. P. B., MRI Hypoxia Measurements. Functional Imaging in Oncology, Springer Berlin Heidelberg: 2014.

2. Hallac, R. R.; Zhou, H. L.; Pidikiti, R.; Song, K.; Stojadinovic, S.; Zhao, D. W.; Solberg, T.; Peschke, P.; Mason, R. P., Correlations of Noninvasive BOLD and TOLD MRI with pO2 and Relevance to Tumor Radiation Response. Magnet Reson Med 2014, 71 (5), 1863-1873.

3. Beeman, S. C.; Shui, Y. B.; Perez-Torres, C. J.; Engelbach, J. A.; Ackerman, J. J. H.; Garbow, J. R., O2-Sensitive MRI Distinguishes Brain Tumor Versus Radiation Necrosis in Murine Models. Magnet Reson Med 2016, 75 (6), 2442-2447.

4. O'Connor, J. P. B.; Boult, J. K. R.; Jamin, Y.; Babur, M.; Finegan, K. G.; Williams, K. J.; Little, R. A.; Jackson, A.; Parker, G. J. M.; Reynolds, A. R.; Waterton, J. C.; Robinson, S. P., Oxygen-Enhanced MRI Accurately Identifies, Quantifies, and Maps Tumor Hypoxia in Preclinical Cancer Models. Cancer Res 2016, 76 (4), 787-795.


Figure 1. (a) R2*-based responsiveness map in a 13762NF tumor and (b) the corresponding H&E staining in the same tumor. The necrotic core (mucinous liquid) was lost during the preparation of the histology slice. Responsive area is located at tumor periphery, which mainly consists of viable tumor as shown by histology.

Figure 2. Dynamic time course of ∆R2* and R1-weighted (TOLD) SI for a progressive hypoxic gas breathing challenge (air to 16% and 14% O2). Baseline is defined as the final 60% data points (in this case, 12 data points) in the air breathing period (20 data points in total). ∆R2* is defined as R2*−R2*(baseline mean). R1-weighted SI is expressed as percent of mean signal intensity at baseline. Change in both R2* and R1-weighted SI in the responsive area of tumor is larger than that in muscle. R1 maps were collected where data points are absent (empty line segments).

Figure 3. Mean ∆R2* and TOLD SI in each breathing period from the experiment shown in Fig 2. ∆R2* and TOLD SI are defined in Fig 2. Three breathing periods (16% O2, first 20 min at 14% O2, and second 20 min at 14% O2) are marked for tumor data. Mean values are calculated by averaging the final 60% data points in each breathing period in each voxel and then averaging across all voxels in a selected ROI (muscle or tumor responsive area). Equation for modified TOLD SI: SI(modified) = SI(raw) / exp(−R2*∙TE), where TE = 5 ms.

Figure 4. Relationship between R1-weighted (TOLD) SI and tumor size for hypoxic gas breathing. TOLD SI is modified for R2* relaxation as described in Fig 3. Increase in R1-weighted signal was observed in tumors larger than 0.4 cm3 for 16% O2 breathing. Small tumors (~0.1 cm3) are associated with decreased R1-weighted signal for the same gas breathing challenge.

Figure 5. (a) Summary of tumor BOLD and TOLD responses with respect to hypoxic and hyperoxic gas breathing. Without modifying for R2*-weighted signal decay, raw TOLD SI decreases during hypoxic gas breathing for all tumors. Employing the modification as described in Fig 3, TOLD SI increases for relatively larger tumors. (b) Illustration of proposed schemes for the observed BOLD and TOLD responses. R2* is dependent on blood deoxyHb fraction and thus increases (decreases) when breathing hypoxic (hyperoxic) gas. For hypoxic gas breathing, R1-weighted SI can be affected by decreased tissue pO2 and increased blood deoxyHb towards opposite directions.

Proc. Intl. Soc. Mag. Reson. Med. 25 (2017)