Combined PET-CT and MRI for response evaluation in patients with squamous cell anal carcinoma treated with curative-intent chemoradiotherapy

Consecutive patients with histologically proven ASCC undergoing baseline and response assessment MRI and FDG PET-CT at a single large tertiary referral centre between January 2018 and April 2020 were identified retrospectively using a keyword search of the institutional Radiology Information System (RIS) (CRIS, Wellbeing Software). Only patients treated with curative-intent CRT using standardised departmental protocols were included. Patients received intensity-modulated radiotherapy with concurrent mitomycin C and 5-flurouracil or capecitabine. In general, patients with T1/T2 N0 disease received 50.4 gray (Gy) in 28 fractions and those with T3/T4 N0 or node-positive disease received 53.2 Gy in 28 fractions to gross tumour volume, using standard protocols. Exclusions were for patients with small surgically resected tumours, patients with metastatic disease not encompassable within the radiotherapy volume or those unfit for curative-intent CRT. The institutional Electronic Patient Record system (EPR) (PPM+) was utilised to obtain patient demographics, clinical history, staging and treatment details. The pertinent follow-up information included PFS and OS. PFS was defined as time from treatment completion to locoregional failure, new distant metastatic disease or ASCC-related death. OS was defined as time from CRT completion to death due to any cause.

Prospective consent for imaging data usage in research and service development projects was obtained from all patients at the time of imaging. Formal ethics committee approval was waived for this study which was considered by the institutional review board to represent evaluation of a routine clinical service.

All PET-CT studies were performed using standardised departmental protocols on Discovery 690 and 710 scanners (GE Healthcare). Both machines used iterative reconstruction, CT for attenuation correction and anatomical localisation, applied scatter and randoms correction. Image reconstruction and acquisition parameters are outlined in Supplemental Table 1. Serum blood glucose was routinely measured prior to imaging and if > 10 mmol/L scanning was not performed. Patients fasted for 6 h prior to intravenous injection of 4 MBq/kg of fluorine-18 FDG. Imaging was acquired from skull base to upper thighs 60 min following tracer injection. No iodinated contrast media was administered.

All MRI scans were performed on a 1.5-T scanner (Siemens Aera, Siemens Healthcare) using a phased array pelvic body coil. The following sequences were acquired: small field-of-view high-resolution sagittal, axial and coronal T2-weighted (T2W) sequences (slice thickness 3 mm); multi-shot turbo spin echo (TSE) T2W sequence (BLADE) (slice thickness 5 mm); axial diffusion-weighted image (DWI) sequences (b50, b700) and associated apparent diffusion coefficient (ADC) map (slice thickness 6 mm). Image reconstruction and acquisition parameters are outlined in Supplemental Table 2. All patients had a contrast-enhanced CT performed as part of initial staging which included a portal venous phase assessment of the liver to exclude metastases.

MRI interpretation was undertaken using the institutional picture archiving and communication system (PACS) (Agfa IMPAX Version 5.6, AGFA Healthcare). PET-CT interpretation was undertaken on a multimodality workstation (Advantage Windows Version 3.5, GE Healthcare). Image analysis was performed by a radiologist with 3 years’ experience under the supervision of a dual-certified radiologist and nuclear medicine physician and a gastrointestinal radiologist with > 15 years’ experience of PET-CT and MRI respectively. All readers were blinded to clinical information prior to interpretation other than the diagnosis of anal cancer. As with routine clinical practice, the post-CRT imaging was compared with baseline studies.

Images were scored on a 5-point scale consisting of complete response (CR), partial response (PR), indeterminate (I), stable disease (SD) and disease progression (PD). Interpretation was applied on a per-patient basis. Discrepant findings were agreed in consensus.

Significantly reduced uptake with reduced tumour volume and resolution of nodal disease with metabolic activity less than mediastinal blood pool activity was categorised as CR; focal uptake greater than mediastinal pool but lower than uptake in the baseline study was categorised as PR; prominent physiological uptake within the anal canal potentially masking residual tumour activity was categorised I; unchanged activity was categorised SD; increased metabolic activity and/or increased tumour volume and/or new nodal disease or metastatic disease was categorised PD.

Tumour response at MRI was assessed on T2W sequences by evaluating morphological appearances and signal intensity within the anorectum and/or nodal disease and/or metastatic disease within the imaged volume. Complete resolution of the primary tumour mass and/or no residual intermediate T2W signal and/or resolution of nodal disease was considered CR; residual intermediate T2W signal intensity was considered PR; areas of abnormality that could not be attributed to residual disease or post-CRT-related changes were considered to be I; increased tumour volume and/or new nodal disease or metastatic disease was categorised PD.

DWI was assessed in conjunction with the corresponding ADC map. High signal on DWI sequences with associated low ADC in the original tumour indicated residual tumour and were considered PR, SD or PD depending on the size of the lesion as per RECIST 1.1 criteria [27].

Concordant findings at both techniques were classified as such. Table 1 illustrates how the overall (combined) criteria were derived when there were divergent criteria at individual PET-CT or MRI assessment.

Physical examination consisting of inguinal lymph node palpation and DRE was carried out 3 months post-CRT. When PET-CT and MRI imaging both demonstrated complete response, clinicians may omit a physical examination in the absence of any symptoms. Standard practice includes a further follow-up MRI and physical examination to be performed at 6 months post-treatment completion. On-going clinical follow-up, including physical examination, is 3-monthly for the first 2 years, 6-monthly years 3 and 4, and 12-monthly in the final year, supplemented by CT scans at 1, 2 and 3 years. True complete responses were defined as cases where locoregional control had been sustained over multiple follow-up evaluations. Evaluation comprised a composite outcome of clinical assessment (history and examination) and imaging follow-up assessment (as described). Additional imaging (e.g. at 9 or 12 months) was requested in equivocal cases, as recommended by the multidisciplinary team (MDT). Any equivocal cases during follow-up combine these assessment modalities to monitor for progression or resolution of changes, and biopsies are reserved for cases with concerning features.

All statistical analyses were performed using SPSS Statistics 26 (IBM Corporation). Two-proportion z test was utilised to compare differences in sensitivity, specificity, PPV, NPV and accuracy. A two-tailed p value of < 0.05 was considered statistically significant. Time to event was measured from date of CRT completion. Kaplan-Meier survival analysis (Mantel-Cox log-rank) Cox regression hazard ratios were calculated and used to assess PFS and OS. Survival curves for OS and PFS were constructed using the Kaplan-Meier method (and compared with the Mantel-Cox log-rank test). Hazard ratios (HR) were calculated using the Cox proportional hazards model.

Seventy-five patients were included; 52 (69.3%) were female and 23 (30.7%) were male with a median age of 62 years (range 35–85 years). The median time interval between end of treatment and MRI was 100 days (range 42–219 days) and 96 days for PET-CT (range 37–255 days). Differences in imaging study scheduling were related to scanner availability, interval unplanned medical care and more recently the COVID-19 pandemic. All patients were followed up either until death or 31^st December 2020 with a median follow-up time of 540 days (range 151–993 days). Table 2 outlines patient characteristics in more detail.

Fifteen of the 75 (20%) patients had persistent anorectal and/or nodal disease after CRT (defined as ‘clinical outcome’). This included 8 patients with residual anorectal disease, 4 with interim progression of the primary tumour and 1 with persistent anorectal and inguinal nodal disease. One patient developed locoregional metastatic disease, and another had new systemic metastatic disease identified on response assessment imaging. Three patients with residual disease (20%) died due to disease progression during the follow-up period, including one individual who progressed through treatment and died shortly after completion of CRT. PFS and OS rates were 79.3% and 95.0% respectively. Median time to progression was 165 days (range 181–557 days), and median time to death was 189 days (range 19–517 days).