Impact of attenuation correction of radiotherapy hardware for positron emission tomography-magnetic resonance in ano-rectal radiotherapy patients

Background: Positron Emission Tomography-Magnetic Resonance (PET-MR) scanners could improve ano-rectal radiotherapy planning through improved Gross Tumour Volume (GTV) delineation and enabling dose painting strategies using metabolic measurements. This requires accurate quantitative PET images acquired in the radiotherapy treatment position.

Purpose: This study aimed to evaluate the impact on GTV delineation and metabolic parameter measurement of using novel Attenuation Correction (AC) maps that included the radiotherapy flat couch, coil bridge and anterior coil to see if they were necessary.

Methods: Seventeen ano-rectal radiotherapy patients received a ${}^{18}F$ -FluoroDeoxyGlucose PET-MR scan in the radiotherapy position. PET images were reconstructed without ( ${\mathrm{CTAC}}_{\mathrm{std}}$ ) and with ( ${\mathrm{CTAC}}_{\mathrm{cba}}$ ) the radiotherapy hardware included. Both AC maps used the same Computed Tomography image for patient AC. Semi-manual and threshold GTVs were delineated on both PET images, the volumes compared and the Dice coefficient calculated. Metabolic parameters: Standardized Uptake Values ${\mathrm{SUV}}_{\max }$ , ${\mathrm{SUV}}_{\mathrm{mean}}$ and Total Lesion Glycolysis (TLG) were compared using paired t-tests with a Bonferroni corrected significance level of $p=0.05/8=0.006$ .

Results: Differences in semi-manual GTV volumes between ${\mathrm{CTAC}}_{\mathrm{cba}}$ and ${\mathrm{CTAC}}_{\mathrm{std}}$ were approaching statistical significance (difference $-15.9\%\pm 1.6\%$ , $p=0.007$ ), with larger differences in low FDG-avid tumours ( ${\mathrm{SUV}}_{\mathrm{mean}}<8.5g{\mathrm{mL}}^{-1}$ ). The ${\mathrm{CTAC}}_{\mathrm{cba}}$ and ${\mathrm{CTAC}}_{\mathrm{std}}$ GTVs were concordant with Dice coefficients $0.89\pm 0.01$ (manual) and $0.98\pm 0.00$ (threshold). Metabolic parameters were significantly different, with ${\mathrm{SUV}}_{\max }$ , ${\mathrm{SUV}}_{\mathrm{mean}}$ and TLG differences of $-11.5\%\pm 0.3\%$ ( $p<0.001$ ), $-11.6\%\pm 0.3\%$ ( $p<0.001$ ) and $-13.7\%\pm 0.6\%$ ( $p=0.003$ ) respectively. The TLG difference resulted in 1/8 rectal cancer patients changing prognosis group, based on literature TLG cut-offs, when using ${\mathrm{CTAC}}_{\mathrm{cba}}$ rather than ${\mathrm{CTAC}}_{\mathrm{std}}$ .

Conclusions: This study suggests that using AC maps with the radiotherapy hardware included is feasible for patient imaging. The impact on tumour delineation was mixed and needs to be evaluated in larger cohorts. However using AC of the radiotherapy hardware is important for situations where accurate metabolic measurements are required, such as dose painting and treatment prognostication.