Estimating Market Size for HER2-Targeted Therapy in 1L Ovarian Cancer: An Epidemiological Funnel Approach

Welcome to our second and final lesson. In our previous session, we established the "who" by precisely defining the DESTINY-Ovarian01 patient population: individuals with newly diagnosed, advanced, HER2-expressing ovarian cancer who are not eligible for PARP inhibitor maintenance therapy. We saw how this strategically positions a HER2-targeted therapy in a segment with high unmet need.

Now, we will tackle the "how many." This lesson focuses on quantifying that opportunity. We will construct an epidemiological funnel, a core tool in market analytics, to estimate the size of the addressable patient population. We will move step-by-step from the total incidence of ovarian cancer down to the specific group of patients eligible for this new treatment paradigm. Your experience in market access will be valuable here, as we will not only build the model but also critically examine the assumptions and real-world factors that influence its accuracy.

By the end of this lesson, you will be able to build a robust framework for sizing this specific market, identifying the key data inputs and sensitivities involved in such a forecast.

The Epidemiological Funnel: A Framework for Market Sizing

An epidemiological funnel, or patient flow model, is a standard method used in the pharmaceutical industry to estimate the number of patients eligible for a particular therapy. It works by applying a series of sequential filters—representing clinical and demographic criteria—to a broad initial population.

We will build a hypothetical funnel for a specific market, such as the United States, to make the exercise concrete. The principles, however, are applicable globally. We will start with all new cases of ovarian cancer and progressively narrow the population based on the criteria for DESTINY-Ovarian01.

Building the Funnel: Step-by-Step Quantification

Let's construct the funnel. At each step, we will identify the key parameter, cite the data source, perform the calculation, and discuss the associated assumptions.

For our starting point, we'll use a hypothetical but realistic figure for annual incidence in the US, which you would typically source from cancer registries like SEER or organizations like the American Cancer Society. Let's assume 19,710 new cases of ovarian cancer annually.

Step 1: Filter for Epithelial Histology
The trial is specific to epithelial ovarian cancer, which is the most common form.

• Parameter: ~90% of ovarian cancers are epithelial.
• Source: This figure is stated in the "About Ovarian Cancer" section of the Daiichi Sankyo press release on DESTINY-Ovarian01.
• Calculation: $19,710\times 0.90=17,739$ patients.

Step 2: Filter for Advanced Stage (FIGO Stage III/IV)
The trial focuses on patients with advanced disease, as this is where maintenance therapy is most critical.

• Parameter: ~75% of patients are diagnosed at an advanced stage.
• Source & Assumption: The press release notes "the majority are diagnosed at an advanced stage." A figure of 75% is a common and reasonable assumption used in epidemiological models for ovarian cancer, though in a real-world project, you would validate this with the most current literature or claims data for your specific market.
• Calculation: $17,739\times 0.75=13,304$ patients.

This group of ~13,300 patients represents the total newly diagnosed, advanced epithelial ovarian cancer population, which is the starting point for any first-line therapy consideration.

Step 3: Filter for Bevacizumab Use in First-Line
The DESTINY-Ovarian01 trial enrolls patients who have already received and are candidates to continue maintenance with bevacizumab. Therefore, we must account for its real-world uptake.

• Parameter: ~60% of patients with advanced ovarian cancer receive bevacizumab in the first-line setting.
• Source & Assumption: This is a critical assumption that is not available in the provided resources. The uptake of bevacizumab varies significantly by geography, physician practice, and importantly, by the patient's HRD status (it is a primary option for HRD-negative patients). An analytics team would invest significant effort to determine this figure through claims data analysis or primary market research. We will use 60% as a plausible placeholder.
• Calculation: $13,304\times 0.60=7,982$ patients.

Step 4: Filter for PARP Inhibitor Ineligibility
As we established in Lesson 1, this is the core strategic positioning of the trial. The reasons for ineligibility are primarily being HRD-negative but can also include other factors.

• Parameter: ~50% of patients are PARP inhibitor ineligible (largely corresponding to the HRD-negative population).
• Source & Assumption: This aligns with the biomarker distribution we discussed in the previous lesson, where roughly half the ovarian cancer population is HRD-negative.
• Calculation: $7,982\times 0.50=3,991$ patients.

Step 5: Filter for HER2 Expression (IHC 1+, 2+, or 3+)
This is the final biomarker filter to define the target population for T-DXd. Determining the prevalence of HER2 expression is critical, and different sources provide varying estimates, which is a common challenge in forecasting.

One source, the trial press release (LINK), states that HER2 expression is present in "up to 55%" of ovarian cancers. However, academic literature often provides more granular and conservative estimates. Let's examine a table summarizing data across multiple studies.

From this table, we can construct the prevalence for the full DESTINY-Ovarian01 population (IHC 1+, 2+, and 3+):

• HER2-low: 17-24%
• HER2-positive: ~14.5%

Combining these gives a range of approximately 31.5% to 38.5%. Let's use a midpoint of 35% for our funnel. This is more conservative than the "up to 55%" figure but is built from a more detailed breakdown. A robust forecast would model scenarios using low, base, and high estimates.

• Parameter: ~35% of ovarian tumors are HER2-expressing (IHC 1+, 2+, or 3+).
• Source: Derived from the data in the "Incidence of HER2 Expression in Cancer" table (LINK).
• Calculation: $3,991\times 0.35=1,397$ patients.

Funnel Summary and The Addressable Market

Let's summarize our epidemiological funnel.

Funnel Step	Patient Population	Parameter	Resulting Patients
Start	Annual Incidence (US, hypothetical)	-	19,710
Step 1	Epithelial Histology	90%	17,739
Step 2	Advanced Stage (III/IV)	75%	13,304
Step 3	Received 1L Bevacizumab	60%	7,982
Step 4	PARP Inhibitor Ineligible (HRD-neg)	50%	3,991
Step 5	HER2 Expressing (IHC 1+/2+/3+)	35%	~1,400

Based on this model, the annual addressable patient population for a therapy like T-DXd in this specific setting in the US is approximately 1,400 patients. This number represents the "peak year" patient opportunity before accounting for factors like market share and duration of therapy.

Real-World Adjustments and Key Sensitivities

As you know, an initial funnel is just the start. An analytics and insights team would spend considerable time pressure-testing the assumptions and incorporating real-world complexities. Key factors to consider for this forecast include:

• HER2 Testing Rates: Our model implicitly assumes that 100% of PARP-ineligible patients are tested for HER2 and their status is known. In reality, upon launch, testing rates would be low and would need to ramp up. The forecast would need to model this adoption curve for HER2 testing, as patients can't be treated if they aren't identified. This is a major go-to-market dependency.
• Data Triangulation: We saw a discrepancy in HER2 prevalence between a press release ("up to 55%") and academic data (~35%). A robust forecast would triangulate information from all available sources, including published literature, conference abstracts, claims data (if available), and primary market research with physicians to arrive at a defensible assumption.
• Interdependency of Assumptions: Our linear funnel is a simplification. For example, the uptake of bevacizumab is likely higher in the HRD-negative population than in the HRD-positive population. A more sophisticated model might segment by HRD status before applying treatment shares to reflect these nuances.
• Market Share: The funnel calculates the total addressable market. A full revenue forecast would require an additional step to estimate the market share T-DXd could achieve against the standard of care (bevacizumab monotherapy), which would depend on the trial's clinical data (e.g., the magnitude of PFS benefit).

Course Conclusion

In this short, focused course, we have moved from a qualitative understanding of a target patient population to a quantitative estimate of the market size.

In the first lesson, we defined the DESTINY-Ovarian01 population, seeing how it cleverly targets a HER2-expressing subgroup within the PARP-ineligible (largely HRD-negative) segment of first-line ovarian cancer—a clear area of unmet need.

In this final lesson, we translated that definition into a concrete number using an epidemiological funnel. We saw that out of nearly 20,000 annual ovarian cancer cases in a market like the US, the specific addressable population for this therapy is a well-defined niche of approximately 1,400 patients per year.

Most importantly, we've highlighted the critical assumptions—around treatment patterns, testing rates, and biomarker prevalence—that an analytics and insights team must investigate and validate to build a credible forecast. You now have a comprehensive framework for how to approach the strategic and analytical challenge of sizing a biomarker-driven opportunity in oncology.