Assessment of HER2+ Breast Cancer Management at Instituto Oncológico Nacional, Panama

Cancer is a disease with high mortality. New treatments require high levels of investment that must be carefully considered by healthcare systems. In this setting, we describe a methodology for supporting decision-making and analyzing the clinical and economic impact of new innovative targeted therapies in the current management landscape. 21%-34% for adjuvant, neoadjuvant and metastatic stage, respectively (24). In addition, the analysis showed that investment in the metastatic stage could optimize the clinical impact and the cost of each percentage point of 5-year OS in HER2+ management, depending on the molecule chosen. We assessed the innovative targeted therapies, which included a new indication for T-DM1. Real world data (26) indicate that 39% of patients who received neoadjuvant treatment with trastuzumab and a taxane would be candidates for T-DM1 treatment. These results are similar to those reported in the Panamanian literature (27). It should also be noted that pertuzumab and lapatinib, the two molecules analyzed for the metastatic disease stage, have different indications: pertuzumab in rst line treatment and lapatinib after progression on trastuzumab. Its use after T-DM1 was assessed in particular, so investment in both therapies is possible. In this analysis, the clinical impact can be regarded as invariable, while the economic impact depends directly on the prices assumed and are subject to changes, for example through price negotiations or other types of managed entry agreements that could modify the expenditure required to expand access to innovative therapies. For the analysis, we used inputs from still immature studies that could lead to bias in the results for T-DM1 assessment in its adjuvant indication, so we recommend that the results from this studies are updated for future assessments. disease and for disease progression in metastatic stage. The difference in expected survival in the metastatic stage of disease compared to the earlier stages and to international standards is marked. The metastatic stage of disease was seen as the stage with the highest potential for improvement in terms of 5-year OS in both relative and absolute terms, namely 62% and 11% respectively, compared to the 1% and 1% that can be achieved in the adjuvant stage and the 0.4% and 0.4% in the neoadjuvant stage, where pertuzumab is already available. Finally, it should also be noted that investment in earlier disease stages is associated with a higher probability of cure, xed treatment duration, denable costs and a better opportunity to delay progression to metastatic disease, which has the highest impact.


Abstract Background
Cancer is a disease with high mortality. New treatments require high levels of investment that must be carefully considered by healthcare systems. In this setting, we describe a methodology for supporting decision-making and analyzing the clinical and economic impact of new innovative targeted therapies in the current management landscape.

Methods
An MS Excel multivariate model was constructed to assess the clinical and economic impact of current disease management and the estimated impact of potential new targeted therapies.

Results
The Instituto Oncológico Nacional (ION) of Panama currently spends $6m annually on the management of 135 new patients with HER2-positive breast cancer, achieving a 5-year overall survival (OS) of 89%, 86% and 18% in adjuvant, neoadjuvant and metastatic disease, respectively. Currently, targeted therapies other than trastuzumab are funded for metastatic stage disease progression and for initial treatment in the neoadjuvant stage. Investment in adjuvant treatments for adjuvant and neoadjuvant disease, and in the metastatic stage would increase 5-year OS to 90%, 86% and 25% respectively (absolute increase of 1%, 0.5% and 7%, respectively).

Conclusions
The 5-year OS at the ION is close to international standards for early-stage disease. However, there are opportunity areas in metastatic disease, if the use of available innovative targeted therapies increased. Although early-stage patients have the highest cure possibilities and de ned costs, investment in rst-line treatment of metastatic disease delays disease progression and offers a high clinical impact in overall survival, both in relative and absolute terms.

Background
Cancer is a disease often characterized by elevated mortality and high therapeutic costs (1). In the US, it was estimated that in 2020, 30% of newly diagnosed cancers in women were breast cancers (2). The average costs of the rst year of treatment for breast cancer in the US in 2010 accounted for about $86,000 per patient (3). In contrast, in developing countries in Latin America, the average cost of treating breast cancer is $15,000, $11,000, and $10,000 in Brazil, Mexico and Costa Rica, respectively (4).
Breast cancer with HER2 receptor overexpression (HER2+) is of particular interest in this setting, not only because breast cancer is the most frequent type of cancer in women (5), but also because around 18% of cases can be classi ed as HER2+ (6). Furthermore, targeted therapies are available that might be expected to improve clinical outcomes, but whose high costs have delayed their implementation as standard treatment in public healthcare systems, particularly in developing countries.
We therefore designed a study to describe and compute relevant variables for decision-making in the context of funding new targeted therapies in a healthcare system.
Our main goals were to provide tools for decision-making in the management of HER2+ breast cancer and to present expected outcomes of the additional use of targeted therapies by de ning and estimating clinical and economic variables.
Clinical impact was assessed from expected 5-year overall survival (OS) and economic impact according to the annual therapeutic budget. The change in the relationship between these variables prompted by the use of new targeted therapies in disease management and the number of patients who bene tted from increased access to these therapies were assessed.

Methods
The current treatment algorithm was modelled in MS Excel, using the main outcome of 5-year OS and the annual cost of treating all patients. The current treatment algorithm was identi ed by type of patient from a review of clinical practice guidelines (7,8) and secondary research (9). HER2+ patients were characterized by their clinical stage, treatment stage, and hormone receptor expression. The number of patients in each group at the Instituto Oncológico Nacional (ION, English: National Oncological Institute) was then estimated according to the distribution these variables. Finally, the treatment regimens and algorithms emerging from this review were con rmed according to medical experience of physicians treating breast cancer at the ION.
The expected 5-year OS by treatment regimen and patient group was estimated with results from clinical trials and studies identi ed through secondary research (10)(11)(12)(13)(14)(15)(16)(17)(18). The OS from the reviewed sources was reported in different time frames or units, for example: median survival or hazard ratio to another known result. An exponential distribution was used to standardize the OS results according to expected overall survival at 5 years (19). The survival result reported in the study was used to estimate the parameter of the exponential distribution used to extrapolate to OS rate in 5-year OS.
To estimate the economic impact, unitary costs of products were retrieved from public documents (20,21) and the costs of non-pharmacological treatment were con rmed with treating physicians at the ION. The suggested dose for each product was also reviewed in clinical guidelines and in the product approvals (22) of the U.S. Food and Drug Administration (FDA). Total cost of treatment per patient was obtained by multiplying the unitary dose cost by the number of doses (Table 1). Some treatment regimens are administered for longer than 1 year, but total costs were computed according to patient incidence, rather than prevalence. In other words, the annual budget for the treated prevalence was estimated by considering the total cost of treating the incident patients, regardless of the duration.
The impact of new therapies was estimated assuming that all patients in the indication would change from their current treatment to the new one, with the corresponding changes in clinical and economic impact. Impact on disease management was assessed by combining the new clinical and economic impact associated with the new treatment, taking into account the number of patients with each outcome. Finally, the clinical and economic impact that would be obtained with the combined use of targeted therapies in distinct patient groups was calculated to assess all possibilities of clinical improvements.

Results
Epidemiology GLOBOCAN 2018 reported an annual incidence of 1,022 new cases of breast cancer in Panama (23), of which 750 were treated at the ION and 135 were HER2+. These patients were classi ed into three clinical stages (based on primary therapeutic data from the ION) to determine their treatment algorithm: adjuvant stage (TNM classi cation from 0 to IIA); neoadjuvant stage (TNM IIB to IIIC); and metastatic stage (TNM IV). It was estimated that 33%, 44% and 23% patients are adjuvant, neoadjuvant and metastatic at diagnosis, respectively; based on the authors' experience, while two thirds of the patients diagnosed in the adjuvant stage will require dual HER2 blockade. An estimated 60% of HER2+ patients may also present overexpression of hormone receptors (HR+).

Treatment algorithm
The adjuvant treatment algorithm is described in Fig. 1. Patients in this stage initially undergo surgery to remove the tumor, followed by a 3-month course of chemotherapy based on doxorubicin and cyclophosphamide. Taxane use was divided: 80% of patients received weekly paclitaxel and 20% received docetaxel every 3 weeks (Q3W). All patients receiving chemotherapy received 1 year of trastuzumab (18 cycles). HR+ patients also received endocrine therapy, consisting of anastrozole in 75% of cases and 25% of tamoxifen. Finally, 85% of adjuvant patients received 5 weeks of radiotherapy (Fig. 1).
The treatment algorithm for neoadjuvant patients is similar to the adjuvant stage. The main difference is that chemotherapy is administered before surgery and pertuzumab is used as targeted therapy concomitantly with trastuzumab; i.e., both treatments are administered for 4 cycles before surgery and patients complete the year of treatment with trastuzumab. Endocrine and radiotherapy regimens are administered as in the adjuvant stage (Fig. 2).
Finally, in the rst line of treatment, metastatic patients receive 10 cycles of trastuzumab and chemotherapy based only on taxanes: 80% of cases receive docetaxel and the remaining 20% receive paclitaxel. Trastuzumab is contraindicated in about 8% of patients, and this cohort is given 9 cycles of capecitabine instead. In total, 70% of patients progress to a second line of treatment, consisting in 90% of cases of ado-trastuzumab emtansine (T-DM1) for 8.5 cycles during the expected 6 months of treatment, and capecitabine or gemcitabine for an average of 7.5 cycles in the remaining 10% of patients. Overall, 57% of the patients who received a second line (i.e., 40% of total patients) would receive a third line of treatment with capecitabine, and 44% of those patients (i.e., 25% of total patients) would nally progress to a fourth line with gemcitabine. Finally, only 10% of patients in this stage receive endocrine therapy with exemestane or anastrozole for 1 year. This assessment highlights the aggressiveness of metastatic disease, in which current cost per patient is low but the cost per 5-year OS was the highest: on average 4.3 times higher than the earlier stages.

Assessment of innovative targeted therapies
The following 4 innovative targeted therapies with 5 indications not currently used at the ION were assessed: pertuzumab, trastuzumab emtansine (T-DM1), lapatinib, and neratinib. As in the assessment of current management, secondary research and FDA product labels (22) were consulted to identify either OS data or hazard ratio, which were later extrapolated to 5-year OS.
Pertuzumab in combination with trastuzumab and a taxane is indicated in all 3 disease stages and currently used at the ION in patients in the neoadjuvant stage. One of the two indications assessed was an 18-cycle adjuvant treatment in patients at high risk of recurrence, de ned as lymph node-positive or hormone receptor-negative cases (25). Given the patient de nition in the algorithm used to assess current management, high risk was de ned as hormone receptor-negative. The other indication assessed was rst-line treatment in metastatic patients, where pertuzumab is given until disease progression or unacceptable toxicity, expected to be 22 months, equivalent to 31 cycles.
T-DM1 is already used in metastatic patients who previously received trastuzumab and a taxane. The new indication assessed in this study was adjuvant treatment in patients who received neoadjuvant treatment with trastuzumab and a taxane and did not achieve pathological complete response (pCR). It was assumed from real world data (26) that 39% of patients who received neoadjuvant treatment with trastuzumab, pertuzumab and a taxane would not achieve pCR and would be considered candidates for adjuvant treatment with T-DM1 for 14 cycles. The use of T-DM1 in an adjuvant indication has shown a reduction of 51% compared to the standard neoadjuvant use of trastuzumab in distant recurrence as the rst invasive disease event and a hazard ratio of 0.7 for mortality compared to standard treatment with trastuzumab; as the study was not yet su ciently mature to report OS, this hazard ratio value was used to estimate the 5-year OS of T-DM1, yielding a 5-year OS of 95.1% of patients in this indication (28).  It should be clari ed that the impact of pertuzumab in the adjuvant stage is less than 1% because only 26% of the patients would be candidates for this treatment. If the impact on these patients alone is compared, the absolute and relative increase in 5-year OS would be 6.8% and 7.7% respectively, requiring an additional $63,000 per patient.  Table 3, where it can be observed that the neoadjuvant stage has the lowest increase in WTP, while metastatic disease has the highest relative OS increase compared to current management.  Table 3 shows that investment in the neoadjuvant stage has the best cost-bene t ratio. However, this stage already has innovative targeted therapeutic options other than trastuzumab. In absolute terms, increases in OS of +1%, +0.5%, +36% were achieved for the adjuvant, neoadjuvant, and metastatic stages, respectively. This suggests that there is an opportunity for additional investment in the metastatic stage of disease, where other innovative targeted therapies, currently used in second line only, can help delay disease progression.

Lapatinib is indicated in combination with
The impact of potential combinations of the 5 indications was assessed, i.e., the combined use of innovative targeted therapies, each administered as indicated. Combinations of two innovative therapies for the same patient group were deemed invalid and were not assessed. As a result, 23 combinations were assessed and their clinical and economic impact were mapped on a Cartesian plane with the origin representing the clinical and economic impact of current disease management, i.e., annual budget, $6 million and expected 5-year OS, 76% (Fig. 4).
Four main quadrants were identi ed in this plane for analysis, labelled as optimal, savings, no-go and trade-off. An optimal quadrant represents an improvement in clinical outcomes with savings: all treatments in this quadrant improve disease management. At the other extreme, the no-go quadrant describes treatments with worse clinical outcomes and increased costs: from a resource optimization perspective, treatments in this quadrant should not be used.
Finally, the savings and trade-off quadrants for all treatments in which clinical and economic impact have the same sign, mean that better clinical outcomes have additional costs and worse clinical outcomes are offset with savings.

Decision-making variables
The main decision-making variable was change in WTP, i.e., the change in total costs versus total bene t; this variable optimizes investment as it decreases. Graphically, current WTP can be represented in the Cartesian plane of clinical and economic impact, e.g., Fig. 4, with a straight line in which total bene t increases are in constant relationship with economic increases. This helps to assess all the combinations by highlighting the combinations on the right or left side of said line, corresponding to decreases and increases in WTP accordingly. In addition, combinations that are closest to the current WTP line can be observed graphically and are deemed as optimal (Fig. 5).
By comparing the 23 combinations against WTP, we found that all combinations would increase the current cost of each percentage point in 5-year OS. It is interesting to note that the use of lapatinib would have the lowest increase in WTP, both around 3%, but with a relative clinical impact of less than 1%. The next indication with lowest impact in WTP is pertuzumab in the adjuvant indication, with a relative increase of 12%, but it also shows a relative clinical impact less than 1%. On the other hand, the indication with the highest relative clinical impact is pertuzumab in the metastatic indication, which increases 5-year OS relatively in the metastatic stage by 62%, resulting in a 5-year OS improvement in HER2+ management of 4% and increases the WTP by 69% (Fig. 5).

Discussion
In this article we analyzed and showed, from a pharmacological perspective, the clinical and economic impact of the potential use of innovative targeted therapies at the ION in a xed time period, that is, without considering changes in disease course due to the increased use of innovative targeted therapies.
Our analysis of the HER2+ patients showed a 5-year OS of 89%, 86%, and 18% for the adjuvant, neoadjuvant and metastatic stages, respectively. These results highlight the notable difference in the survival prognosis in the metastatic disease stage with respect to the other two stages. Results at the ION for early stages are close to the reported international standards ranges of the 5-year OS, with 93%-98%, 78%-89%, 21%-34% for adjuvant, neoadjuvant and metastatic stage, respectively (24). In addition, the analysis showed that investment in the metastatic stage could optimize the clinical impact and the cost of each percentage point of 5-year OS in HER2+ management, depending on the molecule chosen. We assessed the innovative targeted therapies, which included a new indication for T-DM1. Real world data (26) indicate that 39% of patients who received neoadjuvant treatment with trastuzumab and a taxane would be candidates for T-DM1 treatment. These results are similar to those reported in the Panamanian literature (27). It should also be noted that pertuzumab and lapatinib, the two molecules analyzed for the metastatic disease stage, have different indications: pertuzumab in rst line treatment and lapatinib after progression on trastuzumab. Its use after T-DM1 was assessed in particular, so investment in both therapies is possible. In this analysis, the clinical impact can be regarded as invariable, while the economic impact depends directly on the prices assumed and are subject to changes, for example through price negotiations or other types of managed entry agreements that could modify the expenditure required to expand access to innovative therapies.
For the analysis, we used inputs from still immature studies that could lead to bias in the results for T-DM1 assessment in its adjuvant indication, so we recommend that the results from this studies are updated for future assessments.
Furthermore, critical decision-making should consider the logic used in treating a chronic disease: earlier interventions or treatments that in general prevent disease progression have the additional bene ts of a possibility of cure and de nable costs. Additional lines of research, such as the study of equitable access to innovative targeted therapies and analyses of the social impact of investment by stage, for example, considering productivity costs or patients and family associated with progression can contribute to a better informed decision-making process.

Conclusion
The ION currently invests $6m annually in the management of 135 incident HER2+ breast cancers annually, achieving an average 5-year OS of 69%, with 5year OS for early stages close to the reported international standard. It was found that there is recent investment in innovative targeted therapies for neoadjuvant disease and for disease progression in metastatic stage. The difference in expected survival in the metastatic stage of disease compared to the earlier stages and to international standards is marked. The metastatic stage of disease was seen as the stage with the highest potential for improvement in terms of 5-year OS in both relative and absolute terms, namely 62% and 11% respectively, compared to the 1% and 1% that can be achieved in the adjuvant stage and the 0.4% and 0.4% in the neoadjuvant stage, where pertuzumab is already available. Finally, it should also be noted that investment in earlier disease stages is associated with a higher probability of cure, xed treatment duration, de nable costs and a better opportunity to delay progression to metastatic disease, which has the highest impact.  Contrast of clinical and economic impact of combinations