MRI and PET/FMISO In Assessing Tumor Hypoxia in Patients With Newly Diagnosed Glioblastoma Multiforme

Overall Survival (OS) was evaluated every 3 months through end of the study (up to 5 years). A variety of continuous quantitative (functional) imaging features measuring abnormal tumor vasculature (MRI) and hypoxia (FMISO) were evaluated at baseline for their association with Survival time.

Features include

PET Hypoxia measures:

Peak standardized uptake values (SUVpeak); maximum tumor:blood ratio (T/Bmax); and Hypoxia Volume (HV)

DCE MRI perfusion measures:

Mean/median volume transfer constant for gadolinium between blood plasma and the tissue extravascular extracellular space (ktrans)

DSC MRI tumor vasculature:

Normalized Relative cerebral blood volume (nRCBV); and Cerebral blood flow (CBF)

DWI MRI magnitude of diffusion of water through tissue (cell density):

Apparent diffusion coefficient (ADC) using low and high Gaussian distributions

Disease progression was defined by Macdonald criteria. PFS was evaluated every 3months through the end of study (up to 5yrs), features were measured at baseline.

Quantitative imaging features measuring abnormal tumor vasculature (MRI) and hypoxia (FMISO) were evaluated for their association with TTP (cox model) and to discriminate between responders and non-responders at 6 and 9 mos (PFS6 and PFS9) (logistic) Features include

PET Hypoxia measures:

Peak standardized uptake values (SUVpeak); maximum tumor:blood ratio (T/Bmax); and Hypoxia Volume (HV)

DCE MRI perfusion measures:

Mean/median volume transfer constant for gadolinium between blood plasma and the tissue extravascular extracellular space (ktrans)

DSC MRI tumor vasculature:

Normalized Relative cerebral blood volume (nRCBV); and Cerebral blood flow (CBF)

DWI MRI magnitude of diffusion of water through tissue (cell density):

Apparent diffusion coefficient (ADC) using low and high Gaussian distributions

Reproducibility, defined as the variation of repeated measurements in an experiment performed under the same conditions, will be measured as the within subject coefficient of variation with upper an lower repeatability coefficients (LRC, URC) computed as percents from log-transformed data, per Velaquez, et al (J Nucl Med. 2009 Oct;50(10):1646-54. doi: 10.2967/jnumed.109.063347. Epub 2009 Sep 16. PMID: 19759105 ).

Where Within Subject Coefficient of Variation (wCV) is a percentage defined as wCV(%)=100* (exp( SD[ld]/√2) - 1)

and LRC and URC are calculated as: RC=100 (exp(±1.96 SD[ld]) -1).

here SD[ld] is the standard deviation of the difference of the log-transformed PET measurements. These bounds provide an estimate of the lower and upper bounds of percent change observed between scans for each measurement.

Pearson correlation coefficient will be used to quantify the correlation between T/Bmax, the maximum tissue-to-blood ratio activity value, and T/Cmax, the tissue-to-cerebellum activite value Since T/Cmax does not requiring blood sampling and is image derived, a high correlation would indicate that T/Cmax could be an advantageous surrogate for T/Bmax.

Correlation between MRS markers and MR imaging markers and PET markers of tumor hypoxia

MRS markers include:

NAA/Cho, Cho/Cr, Lac/Cr, and Lac/NAA measured within tumor and at the periphery.

MR imaging markers of vascularity include: CBV, CBF, and ktrans PET tumor hypoxia marker: SUVmax

Disease progression was defined by Macdonald criteria. Survival and Progression were evaluated every 3months and at the end of study (up to 5 years) and time to event evaluated.

The FMISO image data were normalized by the average blood activity to produce pixel level tissue-to-blood ratio (T/B) values for all image slices. And the severity of the hypoxia was determined by the pixel with the maximum T/B value (TBmax).

FMISO SUVpeak was determined as the average SUV from a 1 cm circular ROI centered over the hottest pixel. Since FMISO selectively binds to hypoxic tissues, SUVpeak within a region provides a measure of tumor hypoxia.

The hypoxic volume (HV) was determined as the volume of pixels in the tumor on in the FMISO\\PET with a tumor to blood activity ratio ≥ 1.2.

HV is a measure of the spatial extent of tumor hypoxia (in milliliters)

Apparent Diffusion Coefficient (ADC) measures water diffusion through tissue (mm^2/s). Cerebral infarction leads to diffusion restriction resulting in a low ADC signal in the infarcted area.

A double Gaussian mixed model was fit to the ADC histogram and the mean of the lower and the mean of the higher ADC curves were evaluated

Relative cerebral blood volume (RCBV) maps, computed from the integral of ∆R2*(t), were corrected for leakage effects and normalized to normal appearing white matter (nRCBV); nRCBV provides a measure of tumor vasculature Cerebral blood flow (CBF) maps were was normalized to the mean of the region of interest (ROI) in normal appearing white matter (nCBF); nCBF provides a measure of vascular permeability and perfusion

ktrans is a measure of vascular permeability and reflects the rate of gadolinium moves from plasma to extravascular extracellular space (predominantly though blood flow and capillary leakage), which can be represented by the mean or median rate.

Mean & Median ktrans within subject were computed using a matrix-based linearization method to fit tissue ∆R1(t) to the extended Tofts model.

The mean across subjects is presented below (Mean (Mean-ktrans) and Mean(Median-Ktrans))