Intraoperative tests (RD‑100i OSNA system and Metasin test) for detecting sentinel lymph node metastases in breast cancer

The committee considered the heterogeneity and uncertainty in the studies. The committee heard from the clinical specialists on the committee that histopathology practices can vary between hospitals in the UK. It also heard that, in routine clinical practice, the standard of histopathology may be different to that in the studies. It concluded that the extensive research-standard histopathology described in the studies may not be representative of the histopathology conducted in routine clinical NHS practice because it is very time consuming. The committee also concluded that there is variation in how histopathology is performed across the NHS.

The committee discussed the use of the intraoperative tests in clinical practice. The committee noted that RNA extraction is needed for the Metasin test but not for the RD‑100i OSNA system. It raised concerns about quality assurance when performing RNA extraction in the operating theatre and noted that there were advantages in not having to perform this step with the RD‑100i OSNA system. The committee also discussed the support available for users of the intraoperative tests from the manufacturers and considered that there was more uncertainty about the support for the Metasin test than for the RD‑100i OSNA system. The committee concluded that further evidence was needed to show that the Metasin test could be used effectively in hospitals that had not been involved in the development of the test.

The committee considered that biomedical scientists can carry out testing using the RD‑100i OSNA system or the Metasin test, although a higher level of molecular biology expertise appears to be needed for the Metasin test. This expertise may not be available in all hospitals performing breast surgery, particularly once laboratory services are centralised. The committee noted that a biomedical scientist may need to travel to the surgery site to perform the intraoperative test, which would result in a loss of resources from the laboratory during that time. The committee also considered that there is a shortage of pathologists in the NHS and that using these intraoperative molecular tests may free pathology resources for other uses in the NHS.

The committee considered the unpublished non-peer reviewed evidence for the Metasin test. The committee acknowledged the findings of the 2 draft unpublished studies, but noted that the studies had not been peer reviewed and therefore the results should be interpreted with caution. The committee concluded that the test appeared promising but that there was too much uncertainty associated with the evidence to recommend use of the test in routine NHS practice. The committee was encouraging of further research on this test.

The committee considered the 20% prevalence of sentinel lymph node metastases in the patient population used in the base case for the cost-effectiveness analyses. It heard from the clinical specialists that the prevalence was likely to be higher than 20%, and possibly around 30%. The committee noted the sensitivity analysis, which showed that an increase in the prevalence of sentinel lymph node metastases increased the cost effectiveness of the RD‑100i OSNA system (see section 5.37). Therefore, the committee concluded that using the RD‑100i OSNA system was likely to be more cost effective than the base-case incremental cost-effectiveness ratios (ICERs) suggested.

The committee considered the accuracy of histopathology. The committee heard from a specialist committee member that the sensitivity of histopathology in routine UK practice was likely to be lower than 100%, which was what was assumed in the cost-effectiveness analysis. The committee heard that the main source of inaccuracy in histopathology is tissue allocation bias, but that other sources of inaccuracy are possible, including user variability. The committee noted that it was unlikely that a macrometastasis would be missed if the current histopathology guidelines were followed but that it was possible that micrometastases could be missed. However, the clinical significance of this is not known. The committee concluded that it was very difficult to determine the absolute accuracy of histopathology because of the nature of the technique and the inherent variation from qualitative judgement of different pathologists. However, it was likely to be lower than 100% in routine clinical practice.

The committee also considered the assumption in the modelling that histopathology was 100% accurate following discussions about the accuracy of histopathology in clinical practice. The committee concluded that the accuracy of histopathology in any setting could not be 100% because time and resources did not allow every slice of a node to be analysed for metastases. The committee therefore concluded that use of the RD‑100i OSNA system was likely to be more cost effective than the base-case ICERs suggested.

The committee discussed the option to use the RD‑100i OSNA system to analyse either a whole node, or half a node, with the remaining half used to confirm the results by follow-up histopathology. Clinical specialists on the committee said that whole-node analysis had more benefit than half-node analysis followed by histopathology because there was no risk of tissue allocation bias when the whole node was analysed. The committee heard from the manufacturer of the RD‑100i OSNA system that there is a small risk of test failure and the committee considered the impact of this when all of the node tissue is used in whole-node analysis. The manufacturer stated that tissue lysate samples can be stored for up to 1 month and can be re-tested in the event of test failure. The committee concluded that it was acceptable for half-node analysis with postoperative histopathology confirmation to be used while the RD‑100i OSNA system is being validated locally but recommended that, after validation, whole-node OSNA analysis should be fully implemented in local clinical practice to reduce the risk of tissue allocation bias.

The committee considered the possibility of pre-screening tumours for the expression of CK19 and mammaglobin to reduce the small risk of a false negative result. The committee heard from clinical specialists that pre-screening would use pathology resources that are already limited and that it may not be possible to deliver results during surgery. The committee also heard that the interpretation of pre-screening results is complicated by tumour tissue heterogeneity, which means that this information has limited clinical value. Furthermore, the committee heard that the mammaglobin antibody is difficult to work with and that immunohistochemistry staining for mammaglobin is not routinely available in the NHS. The committee concluded that pre-screening for expression of CK19 and mammaglobin was not likely to take place in routine clinical practice because of the resource restraints on pathology services.

The committee considered the difference in the clinical information obtained from intraoperative whole-node analysis compared with histopathological analysis. The committee discussed whether the absence of sentinel lymph node tissue for histopathological analysis following whole-node analysis with the RD‑100i OSNA system posed any risk to clinical decision-making when deciding on the strategy for systemic adjuvant therapy. The committee also considered the risk of a false positive result, such as the detection of CK19 expression from high-grade lymphoma rather than lymph node metastases from breast cancer. The committee concluded that these risks were very low.

The committee discussed the scenario in which intraoperative analysis of sentinel lymph nodes is not performed and when a second operation is needed following identification of lymph node metastases with histopathology. The committee noted that the value in the economic model for the disutility arising from a second operation was 0.03. The committee heard from the clinical specialists that a second breast operation is technically more difficult because of disrupted tissue structure. It also heard that a second operation had an increased risk of complications compared with breast surgery being done for the first time. Furthermore, the committee heard from a patient expert that patients generally found the prospect of having a second operation very worrying and that the option of not having to have a second operation was an important consideration for patients. The committee therefore considered that the disutility from the second operation was likely to be larger than that assumed in the economic model and the cost effectiveness of intraoperative testing was, as a consequence, likely to be underestimated in the base case. The committee concluded that intraoperative analysis of sentinel lymph nodes had considerable advantages over traditional histopathology testing and had the potential to reduce both clinical complications, and patient anxiety and distress.

The committee considered the impact on patients of not knowing what the extent of their surgery would be before the operation. The committee heard from a patient expert the importance of patients being informed before surgery about the different types of surgery they may have, depending on the results of the intraoperative test. The committee noted that there was strong patient preference for all procedures to be done in a single operation and noted that, in routine clinical practice, the uncertainty about the extent of surgery would be explained to the patient.

The committee considered the cost-effectiveness analyses of the RD‑100i OSNA system. The committee acknowledged that the model contained several assumptions that could potentially increase the uncertainty of the cost-effectiveness analysis. However, the committee considered that it was appropriate to assume a higher prevalence than that used in the base case (see section 6.5), an accuracy value of less than 100% for histopathology (see sections 6.6 and 6.7), and that the technical difficulty and complications associated with a second operation may have been underestimated (see section 6.11). The committee concluded that, taking these considerations together, the RD‑100i OSNA system was likely to be more cost effective than the base-case ICERs suggested and that it was likely that the RD‑100i OSNA system was equally or more cost effective than postoperative histopathology. The committee also concluded that the substantial patient benefits associated with using the RD‑100i OSNA system and the strong patient preference for all procedures to be done in a single operation were not fully captured in the cost-effectiveness analyses. The committee therefore concluded that the RD‑100i OSNA system would represent a cost-effective use of NHS resources if used as an option for detecting sentinel lymph node metastases in people with early invasive breast cancer who have a sentinel lymph node biopsy and in whom axillary lymph node dissection is being considered.

The committee considered the cost of histopathology used in the cost-effectiveness analyses and acknowledged it was higher than it is likely to be current practice. The committee was of the view that, if the economic model used a more realistic cost for histopathology, it would indicate that the RD‑100i OSNA system was less cost effective than the base-case ICERs presented in the diagnostics assessment report. However, given all of the uncertainty in the cost-effectiveness analyses (see section 6.13), the committee concluded that the RD‑100i OSNA system was still likely to be a cost-effective use of NHS resources.

The committee heard from clinical specialists about the recent publication of the Z0011 trial, which reported no improvement in survival after axillary lymph node dissection in women who received a positive result for lymph node metastases. The committee heard from the clinical specialists that there was concern that the Z00011 trial was not definitive and possibly under-powered, and that there was no radiotherapy quality assurance programme to monitor the dose of radiation given. The committee concluded that using intraoperative tests for detecting sentinel lymph node metastases offered substantial benefits to patients in current clinical practice and that these benefits were likely to be useful while the uncertainty in the effectiveness of performing axillary lymph node dissection remains.

The committee considered that a quality assurance scheme for the use of the RD‑100i OSNA system and any other relevant intraoperative tests is needed.

The committee considered that, for the efficient use of intraoperative testing, surgical theatre lists may need to be carefully scheduled and multiple analysers may be needed for sentinel lymph node testing if breast operations occur in parallel. The committee heard from clinical specialists that the sentinel lymph node biopsy can be performed first so that the lymph nodes can be analysed using the RD‑100i OSNA system while the primary tumour is being removed. This prevents the time in surgery being significantly increased by use of an intraoperative test.

Clinical specialists on the committee also informed the committee that fewer sentinel lymph node biopsies may be performed during an operating theatre list to allow time to perform axillary lymph node dissections when the intraoperative test results are positive. However, theatre time is made available in the subsequent weeks because the patients are not returning for a second operation, which would occur if patients had to wait for postoperative histopathology results. The committee concluded that any disruption to theatre lists can be overcome with careful planning and scheduling.