Inter-hospital heterogeneity in the authorizations of off-label cancer drugs: A project by Associazione Italiana di Oncologia Medica

Abstract

Purpose

This project aimed to describe variability of local evaluations of off-label cancer drug requests by submitting a list of identical requests to different Italian centers.

Methods

In October 2024, oncology unit directors listed in the AIOM Libro Bianco were invited to participate, evaluating 11 clinical cases with solid tumors through their internal approval process. Supporting evidence was retrospective in two cases, based on phase II trials in seven, and phase III trials in two. Inter-rater agreement was measured using Fleiss’ kappa.

Results

Seven centers completed evaluations. Two cases were approved by all centers, both with a rare cancer with scant alternatives. The mean number of centers approving the individual requests was 5.36/7 (76.6%) with a poor inter-rater agreement (kappa 0.009, 95% CI -0.120 to 0.138). Single centers approved a mean of 8.4 requests (76.6%). The mean number of centers that approved requests was 5.5 for requests based on retrospective studies, 5.3 for those based on phase II trials, and 5.5 for those based on phase III trials. There was no correlation between estimated cost and chance of approval.

Conclusions

There is a significant inter-hospital variability in off-label oncology drug evaluations, leading to potential disparities in patient’s access to care.

Visual Abstract

Keywords

Off-label AIOM cancer drugs law 94/98 inter-hospital heterogeneity

Introduction

In Italy, in the evolving landscape of oncology, the heterogeneity of tumor types, the urgency of treatment decisions, and the frequent lack of approved therapeutic alternatives often necessitate the use of drugs not reimbursed by the Italian Medicines Agency (AIFA). Timely access to innovative therapies remains a critical concern and various mechanisms for early access to unapproved or not-yet-authorized treatments exist in Italy, such as law 648/1996, law 326/2003 (AIFA Fund 5%), law 94/1998 (off-label use), Ministerial Decree 7/9/2017 (compassionate basis) or law 189/2012 (non-negotiated classification).¹ However, awareness and understanding of these access routes are still limited among healthcare professionals. According to a national survey promoted by the Italian Association of Medical Oncology (AIOM) in 2020, 90% of oncologists reported having direct experience with early access programs despite several procedural challenges.^1,2 Moreover, the AIOM survey revealed that not all Italian oncologists are adequately trained or informed about early access frameworks, raising concerns about disparities in treatment opportunities depending on the awareness of the clinician. Another critical point that emerged from the survey is the need for a strong collaboration between oncologists and hospital pharmacists, which is often underestimated but fundamental for navigating regulatory pathways and ensuring drug availability under exceptional circumstances.^1,2

Within this background, use of off-label drug represents a frequent and complex aspect of oncology practice, encompassing regulatory, ethical, and financial considerations. In Italy, off-label drug use is formally regulated by Law 94/98, also known as the “Di Bella” Law, which specifically permits physicians to prescribe authorized medicinal products for off-label indications, routes, or methods of administration/dosage, provided that there are no valid therapeutic alternatives and that favorable clinical data from at least phase II clinical trials support the proposed use. The prescribing physician must obtain patient consent and ensure the presence of sufficient supporting evidence.

The use in clinical practice of already approved drugs in a manner that does not comply with the Summary of Product Characteristics (SmPC) authorized by the European Medicines Agency (EMA) and AIFA should occur under exceptional circumstances and be tailored to the individual patient. Such prescriptions fall under the direct responsibility of the prescribing physician, ensuring flexibility but also imposing a significant ethical and legal burden on the clinician. Then, the cost of the therapy is not reimbursed by the National Health System, but is in charge of the hospital, of course following a positive opinion by an internal pharmaceutical commission (CFI) or analogous commissions. This use is particularly common in oncology due to several factors, such as the high diversity of cancer subtypes, the practical and ethical difficulties of running clinical trials in all patient populations, the swift circulation of early clinical data, and the often-lengthy approval processes by regulatory authorities. The limitations of these requests primarily stem from their strong dependence on the risk–benefit assessments done by the clinician, approvals at a local hospital level, and the absence of mandatory data collection for off-label treatments. This situation increases the risk of empirical and not reproducible decisions in future similar requests. Therefore, the establishment of shared and accessible databases across all centers - containing information on patient and disease characteristics, drug utilization, and outcomes related to efficacy, safety, and quality of life - would be highly advisable.³

Conversely, law 648/1996 allows the prescription of off-label drugs at the expense of the National Health System (NHS), provided there is solid evidence of efficacy and safety supported by national or international scientific research. However, this is not on an individual patient basis: the inclusion of a drug in this pathway requires approval by AIFA’s Scientific and Economic Committee (CSE) and its listing in a dedicated registry published on the AIFA website. The application can be submitted by the CSE itself, scientific societies, or healthcare institutions, and must include detailed documentation on disease severity, lack of valid alternatives, proposed therapy, costs, and supporting literature. Once approved, the drug is reimbursable for all eligible patients nationally.

A recent survey conducted by Cagnazzo et al.⁴ revealed substantial heterogeneity across Italian healthcare institutions in managing off-label drug requests when the drug is not included in the above mentioned law 648/1996. In most centers, the process is initiated by the prescribing physician alone or in collaboration with a clinical research coordinator. Evaluation is commonly performed by a multidisciplinary team, though the final authorization may fall under various institutional bodies, including hospital management or ad hoc committees. Timelines for evaluation vary widely, from less than a week to more than a month, without any legally mandated deadlines and often lacking formal oversight mechanisms. So, authors concluded that there is an urgent need for an update in legislation to standardize this process.⁴

These findings are consistent with earlier data published by Roila et al.⁵ which highlight the prevalence and diversity of off-label prescribing in Italian oncology. In this multicenter observational study, approximately 19% of all chemotherapy prescriptions were off-label. While many were supported by scientific evidence, a notable fraction involved rare tumors, unapproved drug combinations, or use in different lines of therapy than those for which the drug was registered. These findings underscore both the clinical necessity and the structural inconsistency of off-label practices in Italian cancer care.⁵

The implications of this regulatory and procedural heterogeneity are multifaceted. Clinically, the lack of uniform standards may result in divergent treatment pathways for patients with identical conditions, depending solely on their treating institution. Ethically, it raises concerns about equal access to care and the transparency of medical decision-making. Economically, off-label prescriptions not covered by the NHS place a direct financial burden on hospitals, potentially influencing approval decisions at the same level as clinical considerations. To date, Italy still lacks a coordinated, evidence-based system for managing access to therapies beyond standard indications, a gap that fuels fragmentation and can compromise patient outcomes and equal access to care in the Italian landscape.^6,7

In this context, the aim of our study is to better explore the variability in the evaluation and approval of off-label oncology drug requests across different Italian hospitals by submitting a list of identical clinical cases to multiple centers, in order to assess the consistency of internal decision-making processes and highlight potential disparities in patient’s access to treatment.

Materials and methods

This prospective and simulation-based study was designed to empirically assess the extent of inter-hospital variability in the approval of off-label oncologic treatments in Italy. To achieve this, A.A.V, M.D.M. and M.S. chose and developed a standardized set of 11 clinical cases encompassing various solid tumors, different disease settings and diversified treatment options. Being a simulation-based evaluation no real patients were involved, and participating committees were aware they were assessing virtual standardized cases specifically created for this project.

In October 2024, directors of Oncology units listed in the AIOM “Libro Bianco” were invited via e-mail to participate in this study.

The centers who adhered received a web link with the project material to be submitted for their internal approval. The project material consisted of a summary for the 11 cases including the patient’s clinical history, date of birth, literature supporting the request and an estimate of treatment costs for the hospital. All cases and their details are listed in Table 1 . All patients had an Eastern Cooperative Oncology Group (ECOG) Performance Status (PS) of 0–1 (not limiting the eligibility for the active treatment from a clinical point of view) and did not present clinically significant comorbidities or medical and/or laboratory contraindications to the proposed off-label treatment. Evidence supporting the request was retrospective in two cases (n. 1, 3), phase II in seven cases (n. 2, 6, 7, 8, 9, 10, 11) and phase III in two cases (n. 4, 5). Estimated median cost (for 6 months of treatment) was 16107 EUR (range 409-37932).

Table 1.

Details of the 11 cases submitted to evaluation in different Italian centers.

Cases	Date of Birth	Cancer type and setting	Line of therapy	Off-label cancer drug requested	Therapeutic intent	Evidence supporting the request	Estimated Cost	EMA approval status
1	14/05/2000	18/01/2024 Diagnosis of epithelioid hemangioendothelioma (CAMTA-1 positive) with liver metastases at baseline	I line of metastatic disease	Sirolimus	Disease control	Retrospective study^8,9	1 month cost: €552	No
2	30/11/1969	29/06/23 Brain surgery for glioblastoma IDH-wildtype, BRAF V600E mutated, MGMT unmethylated. 08/07/23 Hypofractionated RT + concomitant temozolomide. From 21/03/24 Second line therapy with fotemustine. 23/05/24 Disease progression	III line of metastatic disease	Dabrafenib + Trametinib	Disease control	Phase II trial¹⁰	1 month cost: €5099	No
3	15/01/1950	01/05/2021 Diagnosis of NMIBC underwent TURB + Bacillus Calmette-Guérin. 1/01/2024 TURB: NMIBC very high-risk. Patient refused radical cystectomy and inclusion in SUNRise-1 trial available at the center	Adjuvant line	Intravesical Gemcitabine + Docetaxel	Curative	Retrospective study¹¹	Cost per dose: €22.75	No
4	27/10/1966	22/09/2022 Diagnosis of lung adenocarcinoma with brain metastases (EGFR exon 19 mutated and PIK3CA exon 10 mutated; ALK negative; PDL1 5%). First line therapy: osimertinib and RT on brain lesions. 11/2023 Stable disease in brain; progression disease in thorax. Re-biopsy showed EGFR activating mutation + exon 20 resistance mutation	II line of metastatic disease	Amivantamab + Carboplatin + Pemetrexed	Disease control	Phase III study¹²	Amivantamab €9,052 to €11,315 first month (weight-dependent); €2829 to €3394 every three weeks from cycle 3. Pemetrexed: ~€52/cycle Carboplatin: ~€36/cycle	No
5	25/05/1947	02/2023 Diagnosis of high-grade serous ovarian cancer IIIC, somatic BRCA wild-type, underwent surgery and carboplatin + paclitaxel + niraparib maintenance until 03/2024. 03/2024 Nodal recurrence (platinum free interval > 12 months)	I line of of metastatic disease	Bevacizumab + Pegilated lyposomal doxorubicin	Disease control	Phase III study¹³	Bevacizumab - 70 kg patient: €434/cycle Pegilated lyposomal doxorubicin 30 mg/m²: ~€290.40/cycle = €2323.20 (6 months)	No
6	27/08/1969	01/2020 Diagnosis of hepatocarcinoma; patient underwent transarterial radioembolization. After disease progression: Sorafenib (first line), Regorafenib (second line), Cabozantinib (third line). 2023 Disease progression (Child Pugh A6, Barcelona Clinic Liver Cancer C)	IV line of metastatic disease	Nivolumab	Disease control	Phase II trial¹⁴	Cost per dose: €1527.32	No
7	30/07/1950	28/07/22 Palliative partial penectomy for penile carcinoma (PD-L1 5%, CPS 10%). From 02/09/2023 to 27/03/23 first-line chemotherapy: Carboplatin-Paclitaxel. 27/06/23 Nodal disease progression: second-line chemotherapy with 5-fluorouracil. 09/2023 Further symptomatic nodal progression	III line of metastatic disease	Atezolizumab	Disease control	Phase II trial¹⁵	Cost per dose: €1690.54	No
8	13/07/1999	01/ 2023 Diagnosis of alveolar soft part sarcoma with lung, liver, bone metastases	I line of of metastatic disease	Atezolizumab	Disease control	Phase II trial^16,17	Cost per dose: €1239.03	No
9	14/08/1960	Patient with locally advanced stage IIIB BRAF wild-type melanoma	Neoadjuvant	Pembrolizumab	Curative	Phase II trial¹⁸	Cost per dose: €1243.23	No
10	28/04/1961	02/12/2022: Cholecystectomy for cholangiocarcinoma IV stage (HER2 SCORE 2+ / FISH not amplified; MSS; N-RAS codon 111 mutated; NGS-RNA: no relevant gene fusions found). From 03/2023 to 09/2023 first line treatment: Cisplatin + Gemcitabine. From 10/2023 to 12/2023 second line therapy: XELOX. From 01/2024 to 03/2024 third line therapy: XELIRI. 03/2024 Nodal and hepatic progression; CPS score >10	IV line of metastatic disease	Nivolumab	Disease control	Phase II trial¹⁹	Cost per dose: €1527.32	No
11	14/10/1966	01/ 2023 Diagnosis of metastatic non small cell lung cancer (adenocarcinoma; HER2-mutated; EGFR, ALK, ROS1, NTRK, BRAF, RET negative; PD-L1 <50%). First line: carboplatin + pemetrexed + pembrolizumab. 09/2023 Liver disease progression	II line of metastatic disease	Trastuzumab deruxtecan	Disease control	Phase II trial²⁰	Patient weight: 70 kg Cost per dose: €4741.52	Yes

CAMTA, Calmodulin-binding transcription activator; IDH, Isocitrate Dehydrogenase; BRAF; B-Raf proto-oncogene; MGMT; O6-methylguanine-methyltransferase; RT, Radiotherapy; NMIBC, Non-Muscle Invasive Bladder Cancer; TURB, Transurethral Resection of Bladder; EGFR, Epidermal Growth Factor Receptor; PIK3CA, Phosphatidylinositol-4,5-Bisphosphate 3-Kinase Catalytic Subunit Alpha; ALK, Anaplastic Lymphoma Kinase; PDL1, Programmed Death-Ligand 1; BRCA, Breast Cancer Gene; CPS, Combined Positive Score; HER2, Human Epidermal Growth Factor Receptor 2; FISH, Fluorescence In Situ Hybridization; MSS, Microsatellite Stability; NGS-RNA, Next-Generation Sequencing of RNA; XELOX, Capecitabine and Oxaliplatin; XELIRI, Capecitabine and Irinotecan; ROS1, c-ros Oncogene 1; NTRK, Neurotrophic Tyrosine Receptor Kinase; RET, Rearranged During Transfection.

The local internal pharmaceutical committee of each hospital subsequently evaluated the 11 cases and prepared a summary table containing the evaluation expressed (favorable or not) and any relevant comments. This documentation was then sent to M.D.M for the collection and the analysis of the data.

Statistical analysis

To assess consistency across centers, inter-rater agreement was evaluated using Fleiss’ kappa, a statistical measure for agreement among multiple raters. Along with the estimate of Fleiss’ kappa, 95% confidence intervals were calculated.

Results

Overall, 13 Italian centers adhered to the project. As of 30 April 2025, 7/13 centers completed evaluations: three from Piedmont, one from Emilia-Romagna, one from Lombardy, one from Abruzzo and one from Sicily. The other six centers were repeatedly solicited but they did not complete the evaluation due to internal difficulties.

Out of the seven participating centers, in six cases the evaluation of requests was performed by an internal pharmaceutical commission or analogous commission, while in one case the request was sent directly from the clinician to the pharmacy and to the hospital health management. In the six cases with a commission, at least one hospital pharmacist was a member. A medical oncologist was present in three of these commissions, while in the remaining three no oncologist participated, although in one case the pharmacist belonged to the oncology department.

There is a summary of the evaluations in Figure 1 . Only two cases (n.1 and n. 8) were approved by all the seven centers, both with a rare cancer with scant alternatives.

Figure 1.

Heatmap matrix of evaluations of cases (n. 1-11) by centres (A–G). Green indicates “use off-label of cancer drug authorized”; red indicates “use off-label of cancer drug not authorized.”

The mean number of centers approving the individual requests was 5.36/7 (76.6%), ranging between three (42.9%) and seven (100%), with a poor inter-rater agreement (kappa 0.009, 95%CI -0.120 to 0.138). Single centers approved a mean of 8.4 requests (76.6%), ranging between four (36.6%) and 11 (100%). Centers where the commission included a medical oncologist approved a mean of eight requests (72.7%), ranging between four (36.6%) and 11 (100%), while centers where the commission did not include a medical oncologist approved a mean of 9.3 requests (84.8%), ranging between six (54.5%) and 11 (100%). The mean number of centers that approved requests was 5.5 (range 4-7) for requests based on retrospective studies, 5.3 (range 3-7) for those based on phase II trials, and 5.5 (range 5-6) for those based on phase III trials. There was no correlation between estimated cost for six months of treatment and chance of approval ( Figure 2 ).

Figure 2.

Correlation between estimated cost of treatment and percentage of centers authorizing treatment.

Discussion

This prospective, simulation-based study provides an interesting picture of the real-world heterogeneity of Italian hospital-based evaluations for off-label cancer drug use. Unlike retrospective analyses, our design, where identical clinical cases were simultaneously submitted to different institutions for internal approval, allows for direct comparison of institutional behaviors under standardized conditions.

Unfortunately only seven out of 13 centers have joined the project: this partial participation could introduce a selection bias. It is plausible that centers with more established or responsive internal approval pathways, greater familiarity with off-label procedures, or higher level of engagement in regulatory discussions were more likely to participate. In contrast, non-responding centers may reflect real-world operational barriers such as administrative complexity, lack of dedicated personnel, or simply time constraints. They might be more hesitant or conservative in their approach to non-standard treatments, which could imply that the true extent of inter-institutional heterogeneity across the country is even greater than what we observed. Nevertheless, the perspectives and behaviors of non-responding institutions remain underrepresented.

Despite the small number of centers (n=7) completing the evaluation, our findings indicate marked variability in approval decisions, suggesting fragmentation in the current approach to off-label access. Inter-rater agreement across institutions was nearly null (Fleiss’ kappa = 0.009), underscoring a concerning level of discordance among experts even when faced with the same clinical evidence and economic parameters. Only two of the 11 cases achieved unanimous approval, both pertaining to rare cancers with limited or no effective alternatives, highlighting the clinical urgency that likely outweighed any regulatory or economic reservations. This low concordance rate raises important questions about the objectivity, transparency, and consistency of local approval processes, and ultimately, the equity of patient access to potentially beneficial treatments.

Notably, we found no clear relationship between the level of clinical evidence supporting the request and the likelihood of approval. Requests based on retrospective data were approved, on average, by 5.5 centers, those supported by phase II trials by 5.3 centers, and those based on phase III trials by 5.5 centers. These findings suggest that the strength of the clinical evidence, usually a cornerstone of rational medical decision-making, may not significantly influence institutional determinations in this context. While phase III studies are traditionally viewed as the gold standard for therapeutic validation, our data imply that real-world institutional decision-making is often guided by other factors, which may include individual clinical judgment, hospital policy, past experience, financial aspects or perception of unmet need.

Moreover, tumor type appeared to play a relevant role: requests concerning ultra-rare tumors with no valid alternatives, such as epithelioid hemangioendothelioma and alveolar soft part sarcoma, were uniformly approved, while requests for more common malignancies with available standard options showed greater variability. This observation suggests that institutional decisions may be driven not only by the strength of evidence but also by the perceived rarity and therapeutic urgency of the disease.

Notably, two cases (n. 8 and 9) were related to indications that later in 2025 were included by AIFA in 648/96 list and indication of case 8 was included in the Italian National List of rare diseases. Case 8 received authorization from all the centers, while case 9 was authorized by 6/7 centers.

Similarly, we did not observe a consistent relationship between the estimated treatment cost and approval likelihood. For instance, the intravesical chemotherapy case, characterized by an extremely low per-dose cost (€22.75), was not universally approved, while more expensive options such as checkpoint inhibitors or antibody-drug conjugates were sometimes accepted even with moderate supporting evidence. This suggests that economic considerations, while undeniably relevant in resource-constrained healthcare systems, may not be the primary driver of off-label approval decisions, or at least are weighed variably by different institutions. Of note, some moderately priced therapies (e.g., sirolimus, neoadjuvant pembrolizumab) received relatively high approval rates, further supporting the absence of a direct cost-access relationship. Besides, like evidenced in a financial analysis by Gordon et al.,²¹ in a universal healthcare system, the financing sources for off-label treatments remain a factor that can influence drug access. So, although our analysis did not demonstrate a formal correlation between estimated treatment costs and the likelihood of approval, it is plausible that local hospital budgets and hidden cost-effectiveness considerations may have implicitly influenced certain decisions. The discretion allowed under Law 94/98, combined with variable institutional resources, creates a context where financial sustainability can subtly affect access.

These findings raise the possibility that institutional decision-making may be influenced more by structural and procedural differences across centers than by objective parameters such as cost or quality of scientific evidence. Factors such as internal committee composition, local availability of the drug, previous institutional experience with similar cases, and even philosophical differences regarding therapeutic innovation likely contribute to the observed inconsistency. Moreover, institutional comfort with legal liability under Law 94/98 may also differ, influencing willingness to authorize off-label use, particularly in borderline cases.

Another layer of complexity concerns the legal and ethical implications of offering an off-label treatment under Law 94/98 when a more appropriate framework may be Law 648/96. Some off-label cases submitted for evaluation arguably fit the criteria for applying for inclusion in the lists of Law 648, where a treatment with adequate supporting literature should be made available through centralized AIFA approval, ensuring reimbursement and national-level equity. In these instances, bypassing the 648 route to proceed under 94/98 may be seen as a stopgap measure that limits equitable access, only benefiting those treated at centers willing to authorize such use. This is particularly problematic when the treatment in question, if justified for one patient, should logically be made available to all similarly situated patients. The lack of sustainability in such discretionary off-label use may ultimately undermine the integrity of the public health system and calls for clearer delineation of regulatory pathways. From a legal and policy perspective, these findings highlight important implications for health equity. In a national health system, access to off-label treatments should not vary by geography or institutional risk tolerance. However, under Law 94/1998, prescribers carry personal liability for off-label prescriptions, which may contribute to institutional or individual risk aversion. This legal gray zone further amplifies disparities in patient access across the country. The establishment of national, centralized decision-making tools or shared clinical pathways could provide consistent standards across hospitals, thereby mitigating inequities in access.

A major strength of this study lies in its prospective and standardized design. By using identical clinical cases across centers, we were able to isolate institutional decision-making behaviors without the confounding factors typically found in retrospective datasets. Furthermore, the inclusion of detailed clinical histories, literature references, and cost estimates allowed for a holistic and realistic simulation of real-world off-label approval processes. The diversity of cases in terms of tumor types, lines of therapy, cost profiles, and levels of supporting evidence adds to the generalizability of our findings and reflects the heterogeneity seen in daily oncologic practice. However, the study is not without limitations. The number of responding centers was limited to seven, potentially restricting the representativeness of the results. Although the seven centers were geographically distributed across Northern, Central, and Southern Italy and included both academic referral hospitals and regional institutions, they cannot be considered fully representative of the entire AIOM membership. Therefore, the findings should be interpreted as exploratory and hypothesis-generating rather than definitive at the national level. In addition, despite the request for centers to follow their usual internal processes as they would for real patients, it is plausible that the “virtual” nature of the submissions led to more lenient or idealized decision-making. Such a bias is conservative, in the sense that it likely underestimates the true degree of variability that would be observed in real-life settings, where practical constraints such as drug availability, administrative burden, or budget caps might further reduce approval rates. Still, the persistence of marked heterogeneity even under simulated conditions points to a systemic issue that merits attention. Moreover, basic agreement statistics (Fleiss’ kappa) were used to evaluate inter-rater variability. We acknowledge that more nuanced statistical approaches—such as stratified kappa based on evidence level or multivariate logistic regression—could provide deeper insights into the determinants of approval decisions. However, given the limited number of participating centers (n=7), such multivariable analyses would lack statistical power and be at high risk of overfitting. Specifically, with only seven independent observations at the institutional level, the degrees of freedom are insufficient to reliably estimate a regression model incorporating multiple explanatory variables (e.g., cost, type of evidence, tumor rarity, committee composition). For these reasons, we opted for a descriptive approach, which we believe remains appropriate for the hypothesis-generating nature of this study. Future research with larger and more representative samples could be designed to support more robust modelling and stratified subgroup analyses. Lastly, we did not systematically capture the reasoning behind each center’s decision, limiting to register the positive or negative authorization, and additional comments or explanations were not mandatory. Although we did not collect structured qualitative data using a standardized format, several centers provided open-ended comments which offer valuable insights into their decision-making processes. A recurring theme was the perceived strength of the clinical evidence, often explicitly referencing whether the study supporting the request was a randomized phase II or III trial. Another commonly cited factor was the clinical context, including the rarity of the tumor, the absence of valid alternatives, and the patient’s age or performance status. For instance, treatments for ultra-rare conditions with no effective standard options, such as epithelioid hemangioendothelioma or alveolar soft part sarcoma, were more likely to be approved. Conversely, denials were probably associated with perceived availability of standard treatments (e.g., cystectomy instead of intravesical chemotherapy), concerns about insufficient data, lack of familiarity with the off-label indication in question or less experience of the center in a specific type of cancer. Overall, these unstructured responses confirm that institutional decisions are shaped by a multifaceted interplay of scientific, clinical, logistical, and even cultural factors. Future studies incorporating structured interviews or decision audits would enrich our understanding of the rationale underpinning off-label approvals and refusals.

In conclusion, our findings underscore the lack of a unified, transparent, and evidence-based national framework for off-label evaluations. To address this challenge, several strategies should be considered. First, centralized or regionalized expert panels, possibly involving medical oncologists (for instance, selected from the AIOM regional board) could offer standardized assessments of common off-label requests, promoting consistency while preserving case-by-case flexibility. Second, a national digital registry of off-label requests and outcomes could provide transparency and allow for retrospective learning.³ In this context, for certain disease settings, there are instances in the literature documenting the systematic off-label use of specific drugs - for example, the off-label administration of immune checkpoint inhibitors in the treatment of solid tumors.^22,23 It is therefore incumbent upon the clinician to conduct a thorough review of the scientific literature to determine whether similar off-label uses have already been explored and documented by other investigators in comparable clinical scenarios. Third, clearer guidelines differentiating the appropriate use of Law 94/98 versus 648/96 are urgently needed to prevent misuse and ensure sustainability. Finally, the implementation of shared decision-making models involving pharmacists, ethics committees, patients associations and clinicians could enhance both the procedural rigor and legitimacy of approval processes.

Conclusions

The current system for evaluating off-label oncology drug use in Italy is highly fragmented, with significant consequences for clinical equity and efficiency. Addressing this issue requires coordinated action at both the regulatory and institutional levels, informed by empirical data and guided by principles of evidence-based medicine and patient-centered care.

Footnotes

Author contributions

A.A.V., M.S. and M.D.M contributed to the study conception and design, performing material preparation, data collection and analysis. Other authors contributed to the local evaluation of the clinical cases. The first draft of the manuscript was written by A.A.V. and M.S. and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: M.D.M. reports honoraria from AstraZeneca, Boehringer Ingelheim, Janssen, Merck Sharp & Dohme (MSD), Novartis, Pfizer, Roche, GlaxoSmithKline, Takeda, Viatris, Eisai, Daiichi Sankio for consultancy or participation to advisory boards and institutional funding for work in clinical trials / contracted research from Beigene, Exelixis, MSD, Pfizer and Roche.

M.T. received speakers’ and consultants’ fee from Astra-Zeneca, Pfizer, Eli-Lilly, BMS, Novartis, Roche, MSD, Boehringer Ingelheim, Takeda, Amgen, Merck, Sanofi, Johnson & Johnson, Pierre Fabre, Regeneron, BeiGene, Daiichi Sankyo and Accord. M.T. received institutional research grants from Astra-Zeneca, Boehringer Ingelheim and Roche. M.T. received travel support from Amgen, Takeda and Roche.

G.D.B. reports honoraria from Roche for consultancy or participation to advisory boards, institutional funding for work in clinical trials / contracted research from Roche and speakers’ fee from Roche, Servier, Ipsen, Sanofi, Astra Zeneca.

Other authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Data availability

Data are available after reasonable request to the corresponding author, M.D.M. (massimo.dimaio@unito.it).

ORCID iDs

Anna Amela Valsecchi

Matilde Scaldaferri

Gianmauro Numico

Nicola Personeni

Massimo Di Maio

References

Accesso precoce al farmaco e uso off-label, https://www.aifa.gov.it/accesso-precoce-uso-off-label (accessed 10 May 2025), AIFA (Agenzia Italiana del Farmaco).

Il sondaggio. AIOM: IL 90% DEGLI ONCOLOGI CHIEDE “IN ANTEPRIMA” UN FARMACO INNOVATIVO, https://www.aiom.it/speciale-covid-19-il-sondaggio-aiom-il-90-degli-oncologi-chiede-in-anteprima-un-farmaco-innovativo/ (accessed 10 May 2025).

Fontanals

Esteve

González

, et al. Real-world treatment outcomes of medicines used in special situations (off-label and compassionate use) in oncology and hematology: A retrospective study from a comprehensive cancer institution. Cancer Med. Aug 2023; 12: 17112-17125.

Cagnazzo

Cenna

Federici

, et al. Off-label use in italy: is there space for harmonization? Ann Res Oncol 2023; 3: 165.

Roila

Ballatori

Labianca

, et al. Off-label prescription of antineoplastic drugs: an Italian prospective, observational, multicenter survey. Tumori 2009; 95: 647-651.

Perrone

de Braud

Labianca

, et al. Tackling off-label use of anticancer drugs. J Clin Oncol 2012; 22: 2800; author reply 2801.

Drogovoz

Starikov

Ivantsyk

, et al. Experience and prospects for the use of off-label drugs in oncology. Exp Oncol 2021; 43: 1-5.

Stacchiotti

Miah

Frezza

, et al. Epithelioid hemangioendothelioma, an ultra-rare cancer: a consensus paper from the community of experts. ESMO Open 2021; 6: 100170.

Engel

Cournoyer

Adams

, et al. A retrospective review of the use of sirolimus for pediatric patients with epithelioid hemangioendothelioma. J Pediatr Hematol Oncol 2020; 42: e826-e829.

10.

Wen

Stein

van den Bent

, et al. Dabrafenib plus trametinib in patients with BRAF(V600E)-mutant low-grade and high-grade glioma (ROAR): a multicentre, open-label, single-arm, phase 2, basket trial. Lancet Oncol 2022; 23: 53-64.

11.

Steinberg

Packiam

Thomas

, et al. Intravesical sequential gemcitabine and docetaxel versus bacillus calmette-guerin (BCG) plus interferon in patients with recurrent non-muscle invasive bladder cancer following a single induction course of BCG. Urol Oncol 2022; 40: 9.e1-9.e7.

12.

Passaro

Wang

, et al. Amivantamab plus chemotherapy with and without lazertinib in EGFR-mutant advanced NSCLC after disease progression on osimertinib: primary results from the phase III MARIPOSA-2 study. Ann Oncol 2024; 35: 77-90.

13.

Pfisterer

Shannon

Baumann

, et al. Bevacizumab and platinum-based combinations for recurrent ovarian cancer: a randomised, open-label, phase 3 trial. Lancet Oncol 2020; 21: 699-709.

14.

El-Khoueiry

Sangro

Yau

, et al. Nivolumab in patients with advanced hepatocellular carcinoma (CheckMate 040): an open-label, non-comparative, phase 1/2 dose escalation and expansion trial. Lancet 2017; 389: 2492-2502.

15.

de Vries

Rafael

Gil-Jimenez

, et al. Atezolizumab with or without radiotherapy for advanced squamous cell carcinoma of the penis (The PERICLES Study): A Phase II Trial. J Clin Oncol 2023; 41: 4872-4880.

16.

Bergsma

Elgawly

Mancuso

, et al. Atezolizumab as the first systemic therapy approved for alveolar soft part sarcoma. Ann Pharmacother 2024; 58: 407-415.

17.

Naqash

Coyne

GHOS

Moore

, et al. Phase II study of atezolizumab in advanced alveolar soft part sarcoma (ASPS). J Clin Oncol 2021; 39: 11519-11519.

18.

Patel

Othus

Chen

, et al. Neoadjuvant-adjuvant or adjuvant-only pembrolizumab in advanced melanoma. N Engl J Med 2023; 388: 813-823.

19.

Kim

Chung

Alese

, et al. A Phase 2 Multi-institutional study of nivolumab for patients with advanced refractory biliary tract cancer. JAMA Oncol 2020; 6: 888-894.

20.

Goto

Kubo

, et al. Trastuzumab deruxtecan in patients with HER2-mutant metastatic non-small-cell lung cancer: primary results from the randomized, phase II DESTINY-Lung02 Trial. J Clin Oncol 2023; 41: 4852-4863.

21.

Gordon

Goldstein

Tadmor

, et al. Factors associated with off-label oncology prescriptions: the role of cost and financing in a universal healthcare system. Front Pharmacol 2021; 12: 754390.

22.

Kandemir

Karaomerlioglu

Yaradilmis

. Off-label use of immune checkpoint inhibitors for the treatment of solid tumors: analysis of a nationwide patient sample. J Cancer Res Clin Oncol 2023; 149: 8655-8662.

23.

Stalker

Cohen

, et al. "Off-label" use of checkpoint inhibitors in patients with negative or unknown PD-L1 status in advanced head and neck cancer. J Natl Compr Canc Netw 2025; 23: 90-96.