Magtrace and Sentimag system for locating sentinel lymph nodes for breast cancer

The EAC included a total of 36 studies which included 18 non-randomised controlled trials, 16 cohort studies, 1 prospective paired comparison study and 1 validation study. The 36 studies included a total of 4,202 people who had procedures in which Magtrace and Sentimag was used. Nine studies were reported in conference abstracts only. For full details of the clinical evidence, see section 4 of the assessment report in the supporting documentation.

The published clinical evidence on Magtrace and Sentimag included several comparators. Five studies compared the dual technique radioisotope with Magtrace and Sentimag, and were considered most relevant to the decision problem. These studies were powered to show non-inferiority compared with the dual technique. Eleven studies compared Magtrace and Sentimag with radioisotope alone. Six studies included both the dual technique and radioisotope only but did not report outcomes separately for each comparator. Fourteen non-comparative studies were included for patient reported outcome measures and adverse events only. Studies comparing Magtrace with blue dye alone were not included because of high false negative rates and known inferiority of blue dye when compared with the dual technique.

The published evidence for the detection rates of Magtrace and Sentimag compared with standard care is based on non-inferiority trials. Twelve studies were statistically powered to show non-inferiority, only 1 of which reported dual technique outcomes exclusively. The detection rates per patient for Magtrace ranged from 89.7% to 100% compared with 83.3% to 100% for radioactive isotope with and without blue dye. Detection rates per patient for malignant lymph nodes for Magtrace and technetium‑99m with and without blue dye were also comparable, with ranges of 91.7% to 100% and 90.8% to 100% respectively.

Six studies noted future MRI of the injection and drainage sites being affected by artefacts for up to 5 years after using Magtrace. Chapman et al. (2021) reported the outcome of MRI in 16 people after Magtrace injection. MRI image quality was impaired in all studies and 5 people had non-diagnostic MRI results. Krischer et al. (2017) evaluated MRI done 42 months after Magtrace injection in a sample of 25 people. Imaging interpretation was not restricted in 12 cases, impaired in 10 cases and not possible in 3 cases because of Magtrace residues. There is no longitudinal evidence to determine the effect of this on future diagnoses or treatment.

Warnberg et al. (2019) assessed skin staining up to 36 months postoperatively. People who had retro-areolar Magtrace injections were more likely to experience skin staining compared with those who had peritumoral Magtrace injection, with 67.3% and 37.8% incidence reported at 3 weeks respectively. These figures decreased to 46.2% and 9.4% at 36 months. Karakatsanis et al. (2017) and Karakatsanis et al. (2016) reported that 39.9% and 35.5% of people having breast conserving surgery presented with skin staining. People having mastectomy rarely experienced skin staining. No evidence was identified which directly compared skin staining outcomes of Magtrace with blue dye. A small number of studies that investigated patient reported outcomes did not identify skin staining as a significant problem for patients. For full details of the adverse events, see section 5.3 of the assessment report in the supporting documentation.

Of the 36 included studies, 18 administered Magtrace intraoperatively or on the day of surgery, 7 injected Magtrace within 1 to 3 days of surgery, 5 included people injected more than 3 days before surgery and 6 did not report injection timing. Clinical experts report that Magtrace is usually injected at a routine clinical visit within 30 days of surgery rather than intraoperatively because this gives an improved visual and magnetic signal during surgery. Karakatsanis et al. (2017) note a higher tracer-specific sentinel lymph node detection rate with preoperative administration of Magtrace compared with perioperative administration, with 95.3% and 86.0% respectively (p=0.031). From the available studies, there is no evidence to support a significantly improved detection rate with earlier Magtrace administration.

The company developed a de-novo cost-minimisation analysis from an NHS perspective, which compared Magtrace and Sentimag with the dual technique. The time horizon of the model is from the time the person attends the hospital for sentinel lymph node biopsy to the end of the procedure. The company received resource use data from 3 NHS trusts without on-site access to radiopharmacy or nuclear medicine. The data stated that 30 minutes of theatre time was lost for 20% of all sentinel lymph node biopsies because of delays in surgery or staff shortages. The company model also assumed that 1 additional sentinel lymph node biopsy procedure could be done each week with Magtrace. These 2 factors were included as opportunity costs in the model by measuring the number of sentinel lymph node biopsy procedures foregone, with only 50% of potential additional procedures being realised. The company's base case showed a cost saving of £105 per person using Magtrace and Sentimag. For full details of the cost evidence, see section 9 of the assessment report in the supporting documentation.

Although the EAC accepted most of the assumptions and parameters in the company's base-case analysis, the model was reformulated into a decision tree to improve the clarity of the clinical pathway and allow probabilistic sensitivity analysis. The EAC included 3 arms to represent the different timings and settings of the tracer injections. This included 1 arm for administration of the radioisotope and blue dye, 1 arm for intraoperative Magtrace injection and 1 arm for Magtrace injected at a separate clinic. The model assumes that intraoperative Magtrace injection needs 20 minutes additional theatre time to allow drainage to the axilla. The EAC did not include opportunity costs associated with a lack of availability of radioisotope but did include an opportunity cost for 1 additional sentinel lymph node biopsy procedure each week. The EAC assumes that the opportunity costs are achieved in 50% of hospitals.

When using the EAC's base-case model, Magtrace and Sentimag remains cost saving by £78.90 per person compared with the dual technique. The EAC did a sensitivity analysis to determine the effect of Magtrace injection timing on the cost of the sentinel lymph node biopsy procedure. When time for intraoperative Magtrace injection was more than 29 minutes, Magtrace was cost incurring. The EAC included additional scenario analyses that explored the cost impact of contraindications for Magtrace and the effect on future MRI. Assuming 1% of people cannot have Magtrace because of contraindications and need the dual technique instead, Magtrace remains cost saving by £78.11 per person. If 1% of people in both arms need future MRI and 5% of those who have had Magtrace then have uninterpretable MRI and need gadolinium-enhanced MRI, Magtrace remains cost saving by £78.82.

The key cost driver in the model was the opportunity cost associated with being able to do more sentinel lymph node biopsy procedures each week. Cost savings were dependent on the number of centres that can realise these opportunity costs in clinical practice. It was assumed that using Magtrace would lead to 1 additional sentinel lymph node biopsy per week because of improved management of operating lists. Based on centres doing 400 biopsies annually and 50% of centres realising the opportunity costs, the threshold at which Magtrace became cost incurring was 0.42 additional procedures each week. If no additional procedures are realised, Magtrace would be cost incurring by £58.17 per procedure. If less than 21% of hospitals using Magtrace can do 1 additional sentinel lymph node biopsy procedure each week, Magtrace would be cost incurring.