The results of this study reveal the reality of FDI of the new oncological drugs in the US since 2000. The results of this study have shown an exceptionally high degree of variability in the selection of FDI in terms of the market size of the targeted cancer type, companies’ past development experience, and risks to be accepted for each cancer type. Since this reflects variations in actual decisions of individual developers choosing the most desirable set of a drug and a cancer type when entering the development, the coordinate axes of plots are likely to represent the causes and/or consequences (e.g., a company’s intetion and strategy) of what is considered necessary in the decision.

The main objective of this study was to search for characteristics commonly found in firms’ FDI selection behavior. It is interesting to note that despite the wide variety of projects observed, a common feature of cancer type choice among all projects is the tendency to choose cancer types that have been developed in the past and have a relatively high probability of success.

Our analyses revealed that companies make distinctive first indications based on size (Table 1). For mega-firms, 75 in 102 (65%) projects of solid cancers were focused on major cancer types with high morbidity (e.g., NSCLC, breast cancer, prostate cancer, colorectal cancer, and melanoma). Among hematological cancer projects, 27 in 37 (73%) were for non-Hodgkin lymphoma, CML, and AML. On the other hand, small firms contribute more to projects for rare cancers: 66% of rare solid cancers and 56% rare hematological cancers were developed by small firms.

These differences are also evident in the position of category centroids in the MCA plot. Mega firms tended to enter development with projects targeting larger market-size cancer types with higher mobidity (the horizontal axis in Fig. 2). Smaller firms were more likely to choose cancers other than major cancer types than mega firms. This may underline the justification for interpreting the MCA’s first axis (i.e., market size and/or profit). Given the number of past development projects, it is also not surprising that mega-firms enter development in cancer types in which they have had previous experience compared to smaller firms (the vertical axis in Fig. 2).

It is on the third dimension (the vertical axis of Figs. 3 and 5), which presents the extent of ‘avoiding extremes,’ that the differences in the display of projects between mega and small firms are most clearly evident. The category centroid coordinates of mega and small firm projects are in opposite positions across the origin, and those of large and medium firms fall somewhere between mega and small firms.

On average, the projects of the small firms have an ‘extreme’ profile in terms of development and business risks, while the projects of the mega firms have a ‘moderate’ profile. For example, small firms tend to choose cancer types with very high or very low probability of success, while mega firms tend to choose cancer types in between positions. Mega firms have chosen highly competitive cancer types and moderately competitive areas, while small firms’ choice is skewed toward either highly competitive or less competitive areas. These visualized tendencies are interesting; while many project choices of a small number of mega-firms are similar in targeting relatively moderate profiles, a large number of small firms are highly diverse in their development patterns and attitudes toward risks, each embarking on new projects under very different decision criteria than the mega firms.

The fact that firms of different sizes choose different cancer types is also reflected in the results of the regression analysis (Table 2 and S5). For example, some rare cancer types (e.g., CML, SCLC, and melanoma) were less likely to be chosen by small firms than mega firms. As another example, large and medium firms tended to choose some specific cancer types, including CLL, colorectal, hepatocellular, gastric, and breast cancers. Since only a few firm attributes were available in our analysis, examining the reasons for these imbalances was impossible. However, one possible reason is that firm nationality (or primary target countries/markets) may be confounded in these observations. As observed, projects by Asian companies make up a relatively small percentage of projects by mega and small companies, but about half of projecvts by the medium and large companies. Cancers with a large number of Asian patients (e.g., colorectal cancer, hepatocellular cancer) seem to be chosen more favorably by Asian firms, which could explain some of the observed associations between company size and cancer type.

The results above suggest that firms strategically choose FDI against the background of their respective circumstance; companies with different attributes show different tendency of choices. The next question, then, is whether there are common characteristics that can be seen in the FDI choices of such diverse firms. Our regression analysis revealed three general characteristics common to firms’ choice of cancer types in all first projects. First, firms are more likely to choose cancer types with high development and launch experience; second, cancer types with high five-year survival rates are more likely to be avoided; and third, cancer types with high competition (i.e., many possible entrants) are more likely to be avoided. Although caution must be exercised in interpreting the regression analysis results (‘other conditions being equal’), these characteristics have a certain rationale as the behavior of firms in current new drug development with high uncertainty (i.e., failure risk) and growing costs. They are consistent with the results of previous studies.

Many studies support that experience in the same area of development increases the probability of success for new projects and that companies focus on areas of experience [14,15,16,17,18]. Clinical trials for the development of cancer types with high five-year survival rates take a relatively long time, are generally more difficult, and inflate development costs, so, naturally, they are avoided, other things being equal. The general trend in new drug development in the past decade is avoiding highly competitive cancer types and exploring rare new cancer types that have traditionally been underdeveloped.

Our study focused on analyzing the indication that a company would embark on (i.e., FDI). In our dataset 38.6% of the projects were first approved with different indication from FDI, which means that for quite a few projects FAI is different from FDI. This indicates that many factors must be considered when interpreting FAI, including market competition, priorities in resource allocation, and regulatory processes such as approval review. On the other hand, FDI is more directly linked to the most important objective of obtaining clues about drug efficacy and safety. In particular, the analysis of FDI is considered essential when comparing the behavior of firms with only a few (sometimes only one) development projects with that of large firms, as in this study. The results of this study are useful for future and ongoing consideration of the following objectives. The characteristics and distribution of the current drug development projects provide a predictive picture of how the pharmaceutical market will be structured in the near future. It is of concern to any stakeholder whether this is in the desirable direction regarding efficiency and equity in cancer treatment, and our analysis provides a direct basis for such deliberations. From the perspective of appropriate allocation of social resources, if there is over- or under-entry in certain areas, some intervention (e.g., the provision of incentives) may be necessary to address this. The relationship between the choice of cancer type and firm attributes in the first development project, as revealed by this study, provides a starting point for future support to improve the efficiency of new drug development (e.g., portfolio management) and for real-world policies for effective intervention (if necessary) in the development market as a whole.

This study has several limitations. Firstly, although this study identified some of the background factors in the choice of FDI, it was not aimed to identify the overall mechanism of marketing entry or the process of FDI’s decision. There are multidimensional factors behind firms’ market entry (e.g., firms’ objectives and business conditions, demand for anticancer drugs and medical needs, characteristics of drug candidates to be developed, available technologies). The interpretation of the results of pre-clinical studies should be taken into account in that decision. Studies based on industrial analysis models that consider all of these factors are needed to clarify the mechanisms. Such an investigation is necessary but not possible in the framework of this study, which used only oncological drugs as its sample.

Secondly, consideration of the mechanism linking a new drug candidate’s choice with its chosen indication is also necessary. Whether both choices occur concurrently or whether one precedes the other is an important question that needs to be clarified in itself.

Thirdly, this study focuses on the first step of clinical development, but does not follow the final consequences of the observed projects (i.e., obtaining approvals for inclusion on labels). As seen in the examples where FAI and FDI do not match, uncertainties and LCM strategies not explicitly considered in this analysis may also have an impact on FDI decisions.