Metastatic malignant disease of unknown primary origin in adults: diagnosis and management

Gene-expression-based profiling to identify primary tumours in patients with provisional CUP

Recommendation 1.2.2.9 states that gene-expression-based profiling should not be used to identify primary tumours in patients with a provisional diagnosis of CUP. This recommendation was published in 2010 and was based on evidence from 19 studies, which were of low quality. This was due to the studies being designed to report on diagnostic development and to validate tests, so the link to patient benefit had not been established.

An exceptional surveillance review in 2017 assessed evidence on the use of gene-expression-based profiling for the diagnosis of CUP. It was reported to be associated with high rates of accuracy in classifying tumours and as 'showing promise'. Despite this, the evidence base was assessed as not being sufficient to confirm a benefit of gene-expression-based profiling over the immunohistochemistry techniques in recommendation 1.2.2.7.

A surveillance review of the NICE guideline was started in 2020, which included a consultation with topic experts. Topic experts raised the issue that the recommendation on gene-expression-based profiling for the identification of primary tumours was not reflective of clinical practice and an update maybe required. This was paused due to the COVID-19 pandemic.

An external enquiry was received in 2022, which highlighted that recommendation 1.2.2.9 did not reflect current practice and that genetic testing is available within the NHS (National Genomic Test Directory). However, evidence provided as part of the external enquiry and from topic experts is limited to observational studies and technology development studies (see evidence considered in surveillance). These studies report the proportion of patients in which the primary origin of cancer is successfully identified but do not report on the impact that this diagnosis has on patient outcomes. The current level of RCT evidence in this area is limited, but a large RCT (CUPISCO) is due to complete and publish in 2024 (see ongoing research). This is a phase 2 trial for people who receive disease control with initial platinum based therapy for 3 cycles and are then randomised to more platinum or targeted/immunotherapy based on genomic profile. The primary outcome is progression free survival. The results of this trial are seen as directly relevant to the decision on whether this guideline is updated, therefore we will actively monitor the availability of this trial's results.

Topic experts also identified that the term 'gene-expression-based profiling' may not be a true representation of the processes used in the genomic testing that is currently available. They suggested that a more appropriate terminology is 'molecular diagnostic test' or 'molecular profiling' as this reflects the molecular and genetic testing which is available to CUP patients. These molecular diagnostic tests include not only gene-expression-based profiling but also next-generation genetic or genomic profiling for actionable mutations and immuno-oncology profiling. The terminology of 'gene-expression-based profiling' will therefore be changed to 'molecular diagnostic test'.

Molecular diagnostic testing when deciding which treatment to offer patients with confirmed CUP

We will withdraw recommendation 1.3.2.4, which states that gene-expression-based profiling should not be used when deciding which treatment to offer patients with confirmed CUP. Topic experts highlighted the need to remove the 'do not use' statement on gene-expression-based profiling as it is currently being used to guide treatment both inside and outside of trials in the UK. The removal of this recommendation also aligns with the European Society for Medical Oncology (ESMO) guideline, which advises that next-generation-sequencing findings are discussed at a molecular tumour board to guide treatment. These ESMO recommendations are based on studies, which report the success of genomic profiling or next-generation-sequencing to identify the primary origin and inform treatment for CUP. Further studies on whether this results in improved outcomes for CUP patients are required before the evidence for these techniques can be considered in an update of the NICE guideline. The CAPSICO study has potential to provide these results, so this and the other studies described in the ongoing research section will be actively monitored.

Diagnostic tests

Topic experts highlighted advances in the evidence base for MRI, PET/CT and immunohistochemistry since the publication of the NICE guideline in 2010. For example the ESMO guidelines recommend a PET/CT and brain MRI for patients with single-site or oligometastatic CUP. In comparison the NICE guideline only recommends breast MRI for patients with adenocarcinoma involving the axillary nodes (recommendation 1.2.2.4). Similarly the NICE guideline only recommends a PET/CT scan for patients with provisional CUP presenting with cervical lymphadenopathy or extra-cervical presentations (recommendations 1.2.2.5 and 1.2.2.6).

The Royal College of Pathologists published a dataset for histopathological reporting of CUP in 2018. This provides a broader set of markers to be used in immunohistochemistry panels than those provided in the NICE guideline, which recommends cytokeratin 7 (CK7), CK20, thyroid transcription factor-1, placental alkaline phosphatase, oestrogen receptor and PSA (recommendation 1.2.2.7). This dataset is currently being updated and will be monitored for further advances and this will inform the timing of the guideline update. There is an ongoing study on PET/CT scanning for CUP, this is due to complete in 2024 (FAPI-CUP, see the ongoing research section) and is part of the active monitoring planned for this area of recommendations. A search for ongoing studies will be established in this area to identify other relevant studies for active surveillance.

Views of topic experts

We considered the views of topic experts who were recruited to the NICE Centre for Guidelines Expert Advisers Panel to represent their specialty. Responses were received from 14 topic experts on the need to update the NICE guideline, 7 said that an update was required and 7 said an update was not required. The 14 topic experts were made up of 4 consultant radiologists, 3 consultants in palliative medicine, 3 consultant histopathologists, 3 consultant clinical oncologists and a consultant respiratory physician. The responses to the questionnaire have been grouped into recommendation areas:

Organisation of services

Concern was raised by 1 topic expert over the concept of a CUP team at the hospital or cancer-centre level, as recommended in section 1.1.1. They suggested a potential alternative could be that 1 team or department take the lead role for each site, this would be decided locally, an example could be the Upper-Gastrointestinal tract team as they are possibly already the most diverse staff group. It was considered that this would make administrative and clinical ownership less costly, but could have other advantages such as in reducing delay.

It was reported by 1 topic expert that over 55% of CUP patients present at A&E (no reference provided). They said that a clear directive would be beneficial to emphasise the working relationship between the CUP team and the A&E team in the guideline. This could also consider the role of the GP and how the NICE guideline on metastatic malignant disease of unknown primary origin links to the NICE guideline on recognition and referral of suspected cancer.

Diagnosis

Six of the topic experts stated that the diagnosis section of the NICE guideline needed updating. One topic expert said that a review of all tests included in the section was needed. Two topic experts suggested that the NICE guideline should be updated to reflect the Royal College of Pathologists guidance, educational information and reporting proformas for histopathological reporting of CUP, published in 2018. The gene expression-based-profiling (recommendation 1.2.2.9) was highlighted as being out of date by 2 topic experts.

One topic expert highlighted emerging evidence on the role of whole-body MRI and PET/MRI which could provide a second diagnostic phase (no reference was provided). MRI is currently only recommended for patients with breast cancer involving the axillary nodes (recommendation 1.2.2.4). Similarly PET-CT scans are currently only recommended for patients with specific presentations of extra-cervical or cervical lymphadenopathy with no primary tumour identified on ear, nose and throat panendoscopy (recommendations 1.2.2.5 and 1.2.2.6). This highlights the need for active surveillance in this diagnostic area.

A topic expert suggested that the evidence base for FDG-PET scanning needs reviewing as it may lead to an increase in diagnostic accuracy, reduced delays and a lower number of investigations needed. Although it was highlighted by a different topic expert that a lack of availability of FDG-PET scanning and significant cost may be a potential barrier to implementation in many NHS hospitals.

The need to include recommendations on somatostatin receptor scintigraphy (also called 'octreotide scan') for neuroendocrine tumours was suggested by 1 topic expert.

A topic expert commented that Whole-Body-Diffusion Weighted Imaging (WB-DWI) should be investigated as a possible addition to the recommendations, but did note that significant further research may be needed.

Treatment

One topic expert identified that a large amount of research is ongoing in personalised therapies. They suggested that the use of biomarkers to give personalised treatments for specific tumours and cancers needed to be accounted for in the recommendations. The use of circulating tumour DNA as markers of tumour recurrence and behaviour also has a growing evidence base that could impact on the guideline. This is relevant to recommendation 1.3.2.4.

A review of all treatments used to manage CUP was proposed by 1 topic expert as they thought that new evidence was likely in this area, although no references were provided. There was also a suggestion of a flowchart to inform the clinical management or a prognosticator scheme for patients presenting with CUP.

General

Two topic experts identified guidelines that have been updated recently, ESMO's CUP: Clinical practice guideline for diagnosis, treatment and follow-up (2022) and the American Cancer Society's guidance on the causes, risk factors, prevention, early detection, diagnosis, staging and treatment of CUP (2018). Guideline development and healthcare systems differ and as such European or US guidance may not be directly applicable to the UK setting. The topic experts commented that these include diagnostic and treatment regimens not detailed in the NICE guideline. Examples of the recommendations included by ESMO and not detailed in the NICE guidelines are provided below. The ESMO guideline provides the level of evidence that has informed their recommendations. This applies an adapted version of the Infectious Diseases Society of America - United States Public Health Service Grading System, which ranges from level 1: large RCTs of good methodological quality, to level 5: case reports or studies with no control group. This highlights that many of the ESMO recommendations are based on low quality evidence. The ESMO guideline recommends that:

There are also recommendations based on higher quality evidence in the ESMO guideline that do not exist in the NICE guideline, 2 examples are:

Molecular diagnostic tests (gene-expression-based profiling) to identify primary tumours

We asked topic experts if they agreed that the evidence base on molecular diagnostic tests (gene-expression-based profiling) should be monitored, with a view to updating this area as and when 'sufficient' evidence is available. All agreed with this statement. It was noted that ongoing research is likely to influence future management of this area and an update to the guideline once evidence is available will be necessary, timely and helpful. Reference was made to the current paucity of RCT evidence and that the CUPISCO study may provide practice changing evidence. One topic expert confirmed that the ESMO guidelines also make a similar statement: 'The clinical utility of gene expression profiling to help elucidate the likely primary is not currently supported by high-level evidence. Consequently, it is not generally recommended outside of clinical research'.

Following this, topic experts were asked whether they agreed that the wording of recommendation 1.2.2.9 should be changed from 'Do not use gene-expression-based profiling to identify primary tumours in patients with provisional CUP' to 'For information on gene-expression-based profiling to identify primary tumours in patients with provisional CUP, see the NHS Genomic Medicine Service'.

Five topic experts agreed that the wording should be changed and that it is a clinically appropriate update for UK NHS practice at this juncture.

There was however a concern raised by 1 topic expert that this wording may result in inappropriate use of genetic profiling. This focussed on the potential of a 'free for all' scenario to be created unless a clear directive was in place. They suggested a flowchart to demonstrate where the application could be considered or whether it should be restricted to research. It is not considered possible for NICE to develop a flowchart at this time, due to a lack of evidence on the effectiveness of molecular diagnostic testing to identify primary tumours in patients with provisional CUP. This could be considered for the updated guidelines, with the evidence from the CUPISCO trial likely to be crucial in developing a flowchart.

A further concern was raised by 1 topic expert that the labelling of techniques as gene-expression-based profiling may not be the most accurate terminology. They highlighted that in this area of research there are many non-gene-expression-based profiling techniques that aim to identify likely primary sites which are mostly DNA-based and could be used to inform therapeutic options. They suggested a change to 'molecular diagnostic tests' or 'molecular profiling'. The need for this change is confirmed by the language used in the evidence we have considered during this surveillance review. These studies used terms that may not be covered by gene-expression-based profiling, such as: next-generation-sequencing, molecular guided therapy, genomic alterations, molecular analysis and whole-genomic sequencing.

One consideration when making a change is the consistency of terms across NICE guidelines. The term 'gene-expression-based profiling' is not used in any other NICE guideline so consistency in terminology across other guidelines is not seen as a barrier to removing this description. 'Molecular diagnostics' is a term currently used in NICE's guidelines on brain tumours (primary) and brain metastases in over 16s and haematological cancers: improving outcomes. Another potential alternative is 'genomic biomarker-based', this is used in 3 NICE guidelines which all refer to cancer treatment with a link to the NICE topic page on genomic biomarker-based cancer treatments. These are NICE's guidelines on advanced breast cancer: diagnosis and treatment, oesophago-gastric cancer: assessment and management in adults and prostate cancer: diagnosis and management. The recommendation we are considering here is focussed on the identification of tumours rather than treatment, so this description with a link to the genomic biomarker-based treatments page is not relevant to this recommendation. It therefore seems reasonable to use the molecular diagnostic tests terminology for this recommendation.

We will therefore adopt the molecular diagnostic test terminology and incorporate this into the recommendation wording that was agreed by 5 of the 7 topic experts:

1.2.2.9 For information on molecular diagnostic tests to identify primary tumours in patients with provisional CUP, see the NHS Genomic Medicine Service.

Molecular diagnostic tests (gene-expression-based profiling) when deciding on treatment

The topic experts were asked whether they agreed that recommendation 1.3.2.4 should be withdrawn. Six of the 7 topic experts agreed that this recommendation should be withdrawn and that the evidence be monitored to inform when an update will be appropriate. The topic expert that didn't agree with the proposal was concerned this may result in the inappropriate use of genetic profiling, suggesting that clear guidance on when the application could be used would help to avoid a 'free for all' approach. This could be considered in future updates but currently there is no evidence to direct this development.

The topic experts were then asked whether each of the areas of the guideline required an immediate update or could wait until we have sufficient evidence on gene-expression-based profiling, so that a full review can be completed at this later date. Their responses are summarised in the recommendation areas:

Organisation of services and support

All topic experts agreed that the recommendation section on organisation of services and support could be updated when the molecular diagnostic tests evidence becomes available.

Diagnosis

There was agreement from 6 of the 7 topic experts that the recommendations on diagnosis should be updated when the molecular diagnostic tests evidence becomes available. The use of positron emission tomography–computed tomography (PET/CT) was raised by 1 topic expert. They reported that although the evidence base is low quality, PET-CT scans are seen as an appropriate test for CUP by clinicians. Given the increased availability of PET/CT in the NHS it would be reasonable to update the guidance. The search for ongoing studies conducted as part of this surveillance review identified a study that is assessing PET-CT scans in CUP patients. Although it is a cohort study with no active comparison, the results might inform recommendations 1.2.2.5 and 1.2.2.6, so it would be beneficial to perform active surveillance on this area of evidence.

Immunohistochemistry was also identified as an area for ongoing active surveillance by 1 topic expert. It was suggested that markers recommended for diagnosis have increased in number and a broader panel of markers are now available to pathologists. As identified in 2020 by topic experts, a key data source to consider when updating this area would be the Royal College of Pathologists guidance on histopathological reporting of CUP and malignancy of unknown primary origin. This dataset was published in 2018 and a second edition is in preparation.

Whole-body diffusion weighted MRI (WB-DWI) was also identified as an area for update by 1 topic expert. WB-DWI has shown promise in the diagnosis of multiple myeloma and gastric cancer due to a high contrast resolution which may provide a higher diagnostic performance. A case report has shown potential for WB-DWI to be useful in CUP patients, but further research would be required before this could be considered for our recommendations.

As a result of these issues being raised by topic experts it was decided that a search would be developed for these diagnostic areas and the evidence actively monitored.

Factors influencing management decisions

All topic experts agreed that an update of recommendations on factors influencing management decisions could be delayed until the molecular diagnostic tests evidence becomes available. It is anticipated that the new evidence on molecular diagnostics (if found to be effective) would feed into management decisions.

Managing specific presentations

There were comments from 2 topic experts that the ESMO guidance has been updated recently with specific presentations and the NICE guidelines may need to incorporate these. The section on managing specific presentations in the NICE guideline provides detail on presentations that may benefit from radical treatment. The presentations covered are squamous carcinoma involving upper- or mid-neck nodes and the inguinal nodes, adenocarcinoma involving the axillary nodes and solitary metastases. The ESMO guidance provides specific therapies for breast-like, renal-like, ovary-like and colon-like CUP.

The topic experts said that these changes should be considered when the update of the NICE guideline is completed. All topic experts agreed that recommendations on managing specific presentations could be delayed until the molecular diagnostic tests evidence becomes available.

Systemic treatment

All topic experts agreed that the recommendations on systemic treatment should be updated when further results are available on molecular diagnostics and genomic medicine. It was detailed by 1 topic expert that no evidence exists at the moment for new systemic therapy but the CUPISCO study may provide practice changing evidence.

Views of stakeholders

We did not consult with stakeholders. Although we are not updating the guideline, we will be actively monitoring evidence as it publishes on molecular diagnostics, PET/CT, immunohistochemistry and MRI techniques and the impact of new evidence will be evaluated as it publishes.