COrSIcA: Disease Measurement

Anal cancer is a Human Papilloma Virus (HPV) related cancer with approximately 1500 new cases in the UK per year (2016-2018). Although considered an uncommon cancer, incidence rates have increased threefold in the last 30 years. Initial treatment is chemoradiotherapy, with surgical options reserved for early-stage disease, incomplete response to chemoradiotherapy or disease recurrence. Both treatment modalities are associated with substantial short and long-term side effects. It is estimated that over 90 percent of anal cancer cases in the UK are preventable thus avoiding the associated morbidity and mortality.

A number of sub-populations are at increased risk of developing anal cancer and therefore represent potential targets for cancer prevention and early detection strategies. In addition, a recognised precursor lesion, known as anal intra-epithelial neoplasia (AIN) presents further opportunity for early detection and prevention. AIN can be further stratified into benign low-grade (aLSIL) and potentially cancerous high-grade squamous intra-epithelial lesion (aHSIL). AIN has some similarities to the precursor lesion of cervical cancer, known as CIN. While the natural history of CIN is well-quantified, and there is an evidence-based management strategy of screening and treatment of CIN, with reduction in incidence of cervical cancer, this is so far not the case in AIN.

The recently published 'Treatment in Preventing Anal Cancer in Patients with HIV and Anal High-Grade Lesions' trial (ANCHOR), a large, multi-centre, phase III randomised-controlled trial (RCT) of anal HSIL (aHSIL) treatment versus active monitoring without treatment in people living with HIV (PLWH) showed significantly lower risk of squamous cell carcinoma of the anus (SCCA) with aHSIL treatment than with active monitoring. The magnitude of the benefit observed in the treatment arm led to early closure of the trial (on ethical grounds) and allowed treatment to be offered to those in the active monitoring arm. The likely consequence of such significant findings will be a shift towards screening and treatment of aHSIL and an era of AIN treatment trials to determine the optimal approach.

There are multiple treatments for aHSIL broadly categorised as medical (for example, topical imiquimod, 5-FU, and anti-viral agents such as cidofovir) and surgical (laser, infra-red coagulation, electrocautery, and local excision) therapies. Vaccination is a possible third management strategy. The evidence base underpinning such treatments is of poor quality and the optimal approach cannot be defined.

Assessment of changes in disease burden is an important feature in the clinical evaluation of a treatment. In the context of anal intra-epithelial lesions, clinical assessment involves digital anorectal examination combined with either direct ('naked eye') visual inspection of the anal canal and peri-anus (non HRA), and or visualisation under magnification using a technique known as high resolution anoscopy (HRA). A diagnosis of aHSIL is confirmed histologically following biopsy of suspicious lesions.

An initial scoping exercise of AIN trial literature identified 5 RCTs evaluating AIN treatments. Outcomes assessing disease response to treatment (complete response, partial response, recurrence) typically involve before and after treatment evaluation of one or a combination of HRA guided biopsy and assessment of histology and or assessment of cytological changes[5, 6]. Assessments related to either lesion specific (index lesion) changes or overall disease changes (index lesion and potential metachronous lesions). None of the trials assessed treatment response by any other measure of change in disease burden. Two of the studies did utilise alternative measures of disease burden though this was not for the purpose of assessing treatment response, for example, number of index lesions and sum of the greatest diameters as potential predictors of disease clearance and lesion size (≤50% or ≥50% of the anal canal or perianal region) for stratification purposes. Neither paper reported the technique used to undertake these measurements.

There are a number of potential issues with approaches that rely on histological and cytological outcomes. Histological and cytological outcomes simply confirm the presence or absence of aHSIL; or downstaging of disease to aLSIL, in instances where all aHSIL resolves or at the least is downgraded to LSIL treatment benefit can be inferred, there are however instances where treatment benefits may be inferred in the setting of persisting aHSIL. Lesion size is well known to correlate with disease severity in the cervix. Similarly in the ANCHOR trial, a larger lesion size was associated with an increased risk of progression to cancer. Therefore, even in the presence of persisting aHSIL, by reducing lesion size a treatment may infer benefit through reduced disease burden. Attempts have been made to address this issue in two (NCT02059499 and NCT04055142) of three phase 3 RCT's currently recruiting or in development identified following a search of an international clinical trials register. These trials intend to incorporate 'number of quadrants' and 'number of octants' affected by disease in the outcomes respectively. This approach however might not be an accurate reflection of disease response, for instance, in situations where there are multiple lesions across multiple octants, the size of lesion and number of lesions may change but the spread of disease in terms of number of octants occupied remain unchanged; or in situations whereby octants containing multiple lesions, some of which but not all downgrade to aLSIL, will score the same as octants containing aHSIL lesions, none of which have downgraded.

A further problem is the reliance on HRA to guide assessment, HRA requires specialist equipment and is a skilled procedure. Assessors need to be experienced in subtle lesion characteristics, as such skilled assessors are few and far between. A consequence of a lack of trained HRA assessors opens the potential for bias with investigators administering trial interventions also being those that are undertaking the disease assessment, the same problem is noted in a trial utilising high resolution vulvoscopy for the related condition, vulval intraepithelial neoplasia. A lack of trained HRA assessors could also represent a limiting factor for future large-scale trials. Establishing a method of disease assessment that can be undertaken with photographic representation of disease and is suitable for use in non HRA settings offer potential solutions to such problems.

An acceptable, feasible, reliable and reproducible method of disease measurement would add further granularity to outcomes relating to disease assessment. It could be utilised in aHSIL treatment trials to improve the quality of disease response outcomes and in clinical practice as a modality for communicating disease extent or evaluating response to treatment.

We aim to develop a disease measurement instrument capable of describing disease burden such that it can be used to evaluate disease response to treatment in addition to histological and cytological based measurements further strengthening the quality of disease response outcomes. To account for variations in diagnostic practice, the approach will be applicable to both a HRA and non HRA setting. We are unaware of any previous systematic review or comprehensive methodological work in this clinical area.

A meeting of AIN experts will be conducted to determine a longlist of lesion measurement items capable of capturing disease burden. A series of disease assessments will be undertaken in participants known to have aHSIL to assess disease burden using the longlist measurement items. Disease assessments will be scheduled to coincide with the participants routine disease assessment undertaken as part of the participants 'usual' clinical care for aHSIL. Disease burden for each participant will be captured diagrammatically using location and lesion descriptors recommended by consensus based international guidance for the detection of anal cancer precursor lesions. Disease burden for each participant will also be captured photographically. This detail will be captured initially via non HRA approach (usual care) with the addition of HRA once this approach is established. Described lesions will be correlated with histology from biopsies taken as part of routine clinical practice to confirm the presence of aHSIL in the lesions described.

Two trained disease assessors (experienced clinicians familiar with the assessment of anal intraepithelial lesions) will assess disease burden per participant. Disease burden will also be captured photographically. We will undertake disease assessments on 20-30 participants to obtain an acceptable confidence interval around the estimate of reliability for each measurement item and to ensure variation in the spread of disease patterns for aHSIL is captured. By analysing the results of the clinician assessments and digital analysis of the photographic representation of disease burden, we will be able to determine the best performing measurement items which can be used to inform a disease measurement instrument.

To ensure the disease measurement instrument is relevant (for example, adequacy of photographic representation of disease, the measurement items used), a study advisory group comprising experts in aHSIL assessment and management will be assembled to oversee the study.