LTRN US Equity☆

Item 1A. Risk Factors 58

Item 1B. Unresolved Staff Comments. 104

Item 1C. Cybersecurity 104

Item 2. Properties. 104

Item 3. Legal Proceedings. 104

Item 4. Mine Safety Disclosures. 104

PART II

Item 6. Reserved. 106

Item 7A. Quantitative and Qualitative Disclosures About Market Risk. 114

Item 8. Financial Statements and Supplementary Data. F-1

Item 9A. Controls and Procedures. 115

Item 9B. Other Information. 116

PART III

Item 10. Directors, Executive Officers and Corporate Governance. 116

Item 11. Executive Compensation. 116

Item 14. Principal Accountant Fees and Services. 116

PART IV

Item 15. Exhibit and Financial Statement Schedules. 117

i

SPECIAL

NOTE CONCERNING FORWARD-LOOKING STATEMENTS

This

Annual Report on Form 10-K contains forward-looking statements that involve substantial risks and uncertainties. We make such forward-looking

statements pursuant to the safe harbor provisions of the U.S. Private Securities Litigation Reform Act, Section 21E of the Securities

Exchange Act of 1934, as amended, and other federal securities laws. All statements, other than statements of historical fact, contained

in this Annual Report on Form 10-K, including statements regarding our strategy, future preclinical study activities, future clinical

trial activities, future research activities, future financial position, projected costs, prospects, plans and objectives of management,

are forward-looking statements. The words “anticipate,” “believe,” “contemplate,” “could,”

“estimate,” “expect,” “intend,” “seek,” “may,” “might,” “plan,”

“potential,” “predict,” “project,” “target,” “model,” “objective,”

“aim,” “upcoming,” “should,” ‘will,” “would,” or the negative of these words

or other similar expressions are intended to identify forward-looking statements, although not all forward-looking statements contain

these words. Forward-looking statements reflect our current views with respect to future events and are based on assumptions and subject

to risks and uncertainties.

The

forward-looking statements in this Annual Report on Form 10-K include, among other things, statements relating to:

● our strategic plans to advance the development of any of our drug candidates;

● our expectations related to future expenses and expenditures;

ii

We

may not actually achieve the plans, intentions, or expectations disclosed in our forward-looking statements, and you should not place

undue reliance on our forward-looking statements. Actual results or events could differ materially from the plans, intentions, and expectations

disclosed in the forward-looking statements we make. Factors that may cause actual results or events to differ materially from current

plans, intentions, and expectations include, among other things:

These

factors could cause actual results or events to differ materially from the forward-statements that we make. Furthermore, we operate in

a competitive and rapidly changing environment. New risks and uncertainties emerge from time to time, and it is not possible for us to

predict all risks and uncertainties that could have an impact on the forward-looking statements contained in this Annual Report on Form

10-K.

You

should read this Annual Report on Form 10-K and the documents that we file with the Securities and Exchange Commission, or the SEC, with

the understanding that our actual future results may be materially different from what we expect. These forward-looking statements are

based on management’s current expectations. These statements are neither promises nor guarantees, but involve known and unknown

risks, uncertainties and other important factors that may cause our actual results, performance or achievements to be materially different

from any future results, performance or achievements expressed or implied by the forward-looking statements. Factors that may cause actual

results or events to differ materially from current plans, intentions, and expectations include, among other things, those listed under

Part I, Item 1A. “Risk Factors,” Part II, Item 7. “Management’s Discussion and Analysis of Financial Condition

and Results of Operations” and elsewhere in this Annual Report on Form 10-K. Given these uncertainties, you should not rely on

these forward-looking statements as predictions of future events. The forward-looking statements contained in this Annual Report on Form

10-K are made as of the date of this Annual Report on Form 10-K, and we do not assume any obligation to update any forward-looking statements,

whether as a result of new information, future events or otherwise, except as required by applicable law.

In

addition, statements that “we believe” and similar statements reflect our beliefs and opinions on the relevant subject. These

statements are based upon information available to us as of the date of this Annual Report on Form 10-K, and while we believe such information

forms a reasonable basis for such statements, such information may be limited or incomplete. Our statements should not be read to indicate

that we have conducted an exhaustive inquiry into, or review of, all potentially available relevant information. These statements are

inherently uncertain and investors are cautioned not to unduly rely upon these statements.

Unless

the context requires otherwise, references to the “Company,” “Lantern,” “we,” “us,” and

“our” in this Annual Report on Form 10-K refer to Lantern Pharma Inc., a Delaware corporation, and, where appropriate, its

wholly-owned subsidiaries.

iii

RISK

FACTOR SUMMARY

Our

business is subject to numerous risks and uncertainties, including those described in Part I, Item 1A. “Risk Factors” in

this Annual Report on Form 10-K. These risks include, but are not limited to the following:

iv

● We may be at risk of securities class action litigation.

v

PART

I

Item

1. Business

Overview

We

are an artificial intelligence (A.I.) focused company dedicated to developing cancer therapies and transforming the cost, pace, and timeline

of oncology drug discovery and development. Our development portfolio includes three clinical stage oncology focused product candidates

and consists of small molecules that others have tried, but failed, to develop into an approved commercialized drug, as well as new compounds

that we are developing with the assistance of our proprietary A.I. platform and our biomarker driven approach. Our A.I. platform, known

as RADR®, currently includes more than 200 billion data points, and uses big data analytics (combining molecular data,

drug efficacy data, data from historical studies, data from scientific literature, phenotypic data from trials and publications, and

mechanistic pathway data) and machine learning to rapidly uncover biologically relevant genomic signatures correlated to drug response,

and then identify the cancer patients that we believe may benefit most from our compounds. This data-driven, genomically-targeted and

biomarker-driven approach allows us to pursue a transformational drug development strategy that identifies, rescues or develops, and

advances potential small molecule drug candidates at what we believe is a fraction of the time and cost associated with traditional cancer

drug development. On average, our newly developed drug programs have been advanced from initial A.I. insights to first-in-human clinical

trials in 2-3 years and at approximately $1.0-$2.5 million per program.

We

have active clinical programs for our three lead small

molecule drug candidates: LP-300, LP-184, and LP-284. These programs are focused on multiple important

cancer indications, including both solid tumors and blood cancers. We have established a wholly-owned subsidiary, Starlight Therapeutics,

to focus exclusively on the clinical development of our promising opportunities for central nervous system (“CNS”)

and brain cancers, many of which have no effective treatment options. We are also advancing an antibody-drug conjugate (“ADC”)

program focused on developing highly specific ADCs with highly potent drug-payloads.

In January 2026, we introduced withZeta.ai — a generative AI platform purpose-built to empower researchers

and clinicians to accelerate rare cancer research and drug development, dramatically improve research quality, and reduce R&D costs.

withZeta’s multi-agentic architecture combines intelligent orchestration using a combination of proprietary knowledge bases and publicly

available data with autonomous task completion to deliver a true “co-scientist” experience — one that brings the collective

insight of thousands of domain experts, millions of publications, and billions of data points to address some of oncology’s most difficult

challenges and disease subtypes.

In

2025, the FDA cleared two new Phase 1b/2 investigational new drug (IND) applications for LP-184, further expanding our clinical pipeline

opportunities. The first planned LP-184 Phase 1b/2 trial is positioned to evaluate LP-184 in recurrent triple negative breast cancer

(TNBC) patients as both a monotherapy and in combination with the PARP inhibitor olaparib. The second planned LP-184 Phase 1b/2 trial

is positioned to evaluate LP-184 in a biomarker-defined population of non-small cell lung cancer (NSCLC) patients harboring KEAP1 and/or

STK11 mutations with low PD-L1 expression, in combination with the immune checkpoint inhibitors nivolumab and ipilimumab — a population

with high unmet clinical need and a market opportunity estimated to exceed $2 billion annually. Additionally, LP-184 has received FDA

Fast Track Designations for GBM and TNBC, as well as multiple Orphan Drug and Rare Pediatric Disease Designations across various solid

tumor indications.

Our

strategy is to both develop new drug candidates using our RADR® platform and other machine learning driven methodologies,

and to pursue the development of drug candidates that have undergone previous clinical trial testing or that may have been halted in

development or deprioritized because of insufficient clinical trial efficacy or for strategic reasons by the owner or development team

responsible for the compound. Importantly, these historical drug candidates appear to have been well-tolerated in many instances, and

often have considerable data from previous toxicity, tolerability and ADME studies that have been completed. Our dual approach to both

develop de-novo, biomarker-guided drug candidates and “rescue” historical drug candidates by leveraging A.I., recent advances

in genomics, computational biology and cloud computing is emblematic of a new era in drug development that is being driven by data-intensive

approaches meant to de-risk development and accelerate the clinical trial process. In this context, we are working to create a diverse

portfolio of oncology drug candidates for further development towards regulatory and marketing approval with the objective of establishing

a leading A.I. driven, methodology for treating the right patient with the right oncology therapy.

A

key component of our strategy is to target specific cancer patient populations and treatment indications identified by leveraging our

RADR® platform, a proprietary A.I. enabled engine created and owned by us. Our RADR® platform has grown

to encompass more than 200 billion oncology-focused data points across proprietary, collaborative, and public sources, and employs a

library of 200+ advanced machine learning algorithms. During 2025, we have continued to expand the RADR® platform with

several significant new modules, including: (i) an AI-powered ADC development module that identifies targets, payloads, and tumor selectivity

using a multiomic approach; (ii) a combination regimen module trained on 221 clinical trials to predict the activity and efficacy of

DNA-damaging agent and DNA repair inhibitor combinations; and (iii) a blood-brain barrier (BBB) permeability prediction model that can

process up to 100,000 molecules per hour, for which a PCT patent application has been published with a favorable search report indicating

no significant prior art. Lantern’s BBB prediction algorithms currently hold five of the top ten positions on the Therapeutic Data

Commons (TDC) Leaderboard. We plan to commercially release select RADR® AI modules to the broader research and drug development

community to foster collaborative, open-source innovation in oncology.

Scientific

literature offers a definition for “drug rescue” as research involving abandoned small molecules and biologics that have

not been approved by the U.S. Food and Drug Administration (“FDA”). These rescued molecular compounds are often abandoned

by pharmaceutical companies in the drug discovery or preclinical testing phase, typically because they do not prove effective for the

specific use for which they were developed. Some of these compounds may be useful in treating other diseases for which they have not

been tested. See, Hemphill, Thomas A., “The NIH Promotes Drug Repurposing and Rescue,” Research Technology Management,

v. 5, no. 5, pp. 6-8 (2012). Our use of the term “rescue”, “drug rescue”, or “drug rescuing” refers

to, “...a system of developing new uses for chemical and biological entities that previously were investigated in clinical

studies but not further developed or submitted for regulatory approval, or had to be removed from the market for safety reasons.”,

which is a definition we believe is recognized in the drug discovery, drug development and pharmaceutical and biotechnology industries.

See, Naylor, S. and Schonfeld J., “Therapeutic Drug Repurposing, Repositioning and Rescue,” DDW (Drug Discovery World)

Winter 2014, and Mucke, HAM, A New Journal for the Drug Repurposing Community. Drug Repurposing, Rescue & Repositioning 1, 3-4 (2014).

The use of the term “drug rescue,” “rescuing,” or words of similar meaning in this report should not be construed

to mean that our RADR® platform has resolved all issues of safety and/or efficacy for any of our drug candidates. Issues

of safety and efficacy for any drug candidate may only be determined by the U.S. FDA or other applicable regulatory authorities in jurisdictions

outside the United States.

Our

current portfolio consists of three lead drug candidates that are in clinical phases (known as LP-300, LP-184 and LP-284) and an Antibody

Drug Conjugate (ADC) program that is in preclinical research optimization. In January 2023, we formed a wholly owned subsidiary, Starlight

Therapeutics Inc. (“Starlight”), to develop drug candidate LP-184’s central nervous system (CNS) and brain cancer indications

– including glioblastoma (GBM), brain metastases (brain mets.), and several rare pediatric CNS cancers. Following the formation

of Starlight, we may also refer to the molecule LP-184, as it is developed in CNS indications, as “STAR-001”. All of these

drug candidates and our ADC program are leveraging precision oncology, A.I. and genomic driven approaches to accelerate and direct development

efforts.

We

are conducting a targeted phase 2 trial (the HarmonicTM trial) for LP-300 in never smoking

patients with advanced non-small cell lung cancer (“NSCLC”) in combination with chemotherapy, under an existing investigational

new drug application. Our candidate LP-184 has shown promising in-vitro and in

vivoanticancer activity in multiple solid tumor indications (including pancreatic, lung, bladder,

glioblastoma and triple negative breast cancer), and enrollment has now been completed in a Phase 1a clinical trial for LP-184. Based

on the results and insights from the LP-184 Phase 1a clinical trial, we are advancing and optimizing development plans for multiple future

LP-184 clinical studies. Our candidate LP-284 has shown promising in-vitro and

in vivoanticancer activity in multiple hematological cancers, which are distinct

from the indications targeted by LP-184. LP-284 is advancing in a Phase 1A clinical trial.

Our

ADC program has also continued to advance. During 2024 and in 2025, we continued to apply our RADR® A.I. platform to advance

and refine an A.I. powered module focused on improving the precision, cost and timelines of ADC development for cancer. In

2023, we entered into a research collaboration with Bielefeld University in Germany focused on development of ADCs utilizing cryptophycin

as the ADC drug-payload. Cryptophycins are promising antitumor molecules that have demonstrated potency at ultra-low, picomolar, concentrations.

In a broad range of preclinical studies, the cryptophycin-ADC synthesized as part of the Bielefeld

collaboration demonstrated promising picomolar level potency and anti-tumor activity in multiple solid tumor cell lines, including breast,

bladder, colorectal, gastric, pancreatic and ovarian.

In

addition to our lead drug candidates and ADC program, we also have an additional drug candidate, LP-100, that we believe has potential

for future development in combination with the class of anticancer agents known as PARP inhibitors (PARPi). For LP-100, as well as our

lead drug candidate LP-300, we have leveraged data from prior preclinical studies and clinical trials, along with insights generated

from our A.I. platform, to target the types of tumors and patient groups we believe will be most responsive to the drug. Both LP-100

and LP-300 showed promise in important patient subgroups, but failed pivotal Phase 3 trials when the overall results did not meet the

predefined clinical endpoints. We believe that this was due to a lack of biomarker-driven patient stratification.

LP-300

has been studied in multiple randomized, controlled, multi-center non-small cell lung cancer, or NSCLC, trials that included administration

of either paclitaxel and cisplatin and/or docetaxel and cisplatin. LP-100 has previously been in a genomic signature guided phase 2 clinical

trial in Denmark for patients with metastatic castration resistant prostate cancer (mCRPC). 9 patients (out of a targeted enrollment

of 27) were treated in the trial. The median overall survival (OS) for the initial group of 9 patients was approximately 12.5 months,

which is an improvement over other similar fourth-line treatment regimens for mCRPC. Based on our evaluation of the synergies of LP-100

with PARP inhibitors, the decision was made in the first quarter of 2023 to close the phase 2 clinical trial in Denmark, to allow the

focus of LP-100-directed resources on positioning the molecule for development in earlier lines of therapy with potentially larger market

opportunities. LP-100 was previously out-licensed by us to Allarity Therapeutics A/S. In July 2021, we entered into an Asset Purchase

Agreement to reacquire global development and commercialization rights for LP-100 from Allarity.

Our

development strategy is to pursue an increasing number of oncology focused, molecularly targeted therapies where artificial intelligence

and genomic data can help us provide biological insights, reduce the risk associated with development efforts and help clarify potential

patient response. We plan on strategically evaluating these on a program-by-program basis as they advance into clinical development,

either to be done entirely by us or with out-licensing partners to maximize the commercial opportunity and reduce the time it takes to

bring the right drug to the right patient.

As

part of our overall growth strategy, we plan to grow our pipeline by identifying new drug candidates and pursuing potential indications

for LP-300, LP-184, LP-284, our ADC program and other drug candidates while leveraging our RADR® platform. We are also

pursuing the identification and design of potential combination therapies in cancer for our compounds by leveraging our RADR®

platform to analyze synergistic genomic networks and biological pathways with other currently approved drugs.

In addition, in 2026 we plan to introduce our proprietary artificial intelligence (AI) platforms and related technologies

as a potential source of revenue. This initiative is intended to leverage our existing AI infrastructure, technologies and drug development

expertise to create new opportunities in precision oncology and translational research support. We expect to evaluate multiple partnership

and commercialization models as our platforms advance and approach market readiness through both subscriptions, access and services for

biopharma companies, researchers, drug developers, and other users.

We

have an extensive multi-national portfolio of intellectual property directed to our drug candidates, and to protect the targeted use

and development of our portfolio of compounds in specific patient populations and in specific therapeutic indications. In addition, as

our RADR® platform and other machine learning driven methodologies progress and mature, we will continue to evaluate additional

ways to further protect these assets.

As

of March 1, 2026, we own or control over 200 active patents and patent applications across 20 patent families whose claims are directed

to our drug candidates and what we plan to do with our drug candidates. We have in-licensed or acquired patents and patent applications

from AF Chemicals and BioNumerik Pharmaceuticals that are directed to the compounds LP-184, LP-284, LP-100 and LP-300, and methods of

using the compounds. Additionally, we have also filed patent applications to further enhance and extend the use of these compounds. Our

patent families are directed to our drug candidates, their usage, manufacturing and other matters. These matters are essential to precision

oncology and relate to: (a) data-driven, biologically relevant biomarker signatures, (b) patient selection and stratification approaches

that rely on prediction of response derived from these signatures and, (c) the ability to develop novel, combination therapy approaches

with existing therapeutics.

Our

Drug Candidate Pipeline

One

of the ways we built our drug candidate pipeline is by in-licensing clinical stage drug candidates that may have been discontinued for

development. We use our RADR® platform to assist in analyzing prior clinical research conducted by others to identify

small-molecule oncology drug candidates that have (i) a well-tolerated profile evidenced by completion of phase 1 clinical trials, and

(ii) demonstrated at least limited antitumor or anticancer activity in clinical trials. We intend to advance the drug candidates in our

pipeline as potential precision medicine treatments for cancer. Our targeted development workflow includes preclinical studies where

drug activity and associated gene signatures are identified, in part through strategic collaborations with some of the top academic institutions

and clinical translational centers in the world. Using this collaborative approach, together with innovative observations from our RADR®

platform, we intend to develop and add drug candidates for our pipeline with the objective of treating the right patient populations

with the right oncology therapies.

Our

current pipeline of development programs includes our three lead small molecule drug candidates: LP-300, LP-184, and LP-284, and our

Antibody Drug Conjugate (ADC) program.

We

currently have INDs in the U.S. for LP-300, LP-184 and LP-284.

Our

Precision Cancer Therapy Development Using Our Innovative RADR®Platform

We

believe RADR® is one of the world’s largest A.I and machine learning (M.L.) oncology drug discovery and development

platforms, consisting of over 200 billion oncology-focused data points. These data points consist of large-scale multi-omic data, derived

from over 130,000 patient records, over 150 drug-tumor interactions, thousands of drug classes, and covering over 135 cancer subtypes.

RADR® leverages this data and over 200 advanced ML algorithms to power its drug discovery and development modules. RADR®’s

data, capabilities, and insights have powered the development of new Lantern drug candidates, advancement of new indications for existing

drugs, and identification of potential new drug combinations.

Historically,

cancer treatment protocols include surgery, chemotherapy and radiation therapy. Treatments have been selected based on histologic type

and disease spread, irrespective of genetic differences among patients. With the advent of precision therapies, cancer treatments increasingly

target specific genes or mechanisms of action for a more personalized approach to patient care. This trend represents a substantial advance

in cancer treatment because tumor growth is highly dependent on genetic changes and the genetic profile of the individual and the progression

of the disease is highly variable amongst patients.

Our

RADR® platform is core to our drug development approach for identifying the desired candidates to in-license and develop.

Oncology drug development is exceedingly challenging, with an overall estimated Phase 1-to-approval probability of success of just 3.3

percent (according to reports from DIA Global Forum: What Are the Chances of Getting a Cancer Drug Approved?, May 2019)

and an estimated mean cost to deliver a new oncology medicine of $4.4 billion (Study published in Targeted Oncology, 2023; Analysis

of the Cost of Developing Oncology Drugs Approved by the FDA Between 1997 and 2020). There is a critical need to rescue clinical

research on drugs that have failed clinical trials in order to provide additional possible therapies for patients while reducing the

overall cost of therapeutic development. Many drug failures within oncology may be attributed to the heterogeneity of the tested patient

population, even though there may be a strongly positive therapeutic impact on certain patient subgroups within that population.

As

data-centric and machine learning approaches are beginning to change the pace and scale of drug discovery and development, research and

development (“R&D”) we believe efforts in large biopharma companies are beginning to shift away from traditional approaches

towards new data and A.I.-centric approaches. According to Deloitte Consulting, in Ten Years On | Measuring the return from pharmaceutical

innovation 2019, “decades of advances in science and technology have driven improvements in health care outcomes and influenced

stakeholder expectations of the role of the biopharmaceutical industry (biopharma)”. The Deloitte Consulting report further describes

that R&D costs will, “shift from traditional discovery and trial execution to a process driven by large datasets, advanced

computing power and cloud storage”. Continuing the trend of scientific advancement impacting biopharma, as noted in Deloitte’s

2019 report, the findings in Deloitte’s 2023 report show R&D returns rebounding, with regulatory challenges and the need for

AI integration remaining as key future hurdles for the industry.

Analysts

estimate that this shift from traditional screening, and trial-based studies to leveraging in silico, data and A.I. methodologies has

driven a significant increase in the spending on A.I. by the biopharma and drug discovery community. According to GlobalData, the drive

to reduce drug development time and costs through AI-enhanced computer-aided drug design, coupled with a surge of AI-focused startups,

is projected to result in biopharma AI spending reaching $3 billion by 2025. As a result of these trends and changes in the R&D model

in biopharma, we believe that we, and companies that are using data-centric and A.I. centric approaches to drug discovery and development,

are in an ideal position to benefit from this industry shift that has the potential to help deliver drugs to the right patients faster,

with a higher degree of personalization and a potentially lower amount of average costs in the development cycle.

Our

drug rescue approach leverages substantial prior research and development investments in candidates that were withdrawn from development

prior to submission for FDA approval. The large volume of failed compounds, recent developments that permit increased access to validated

genomic and biomarker data, and the rapid evolution of A.I. technology creates an opportunity to efficiently capitalize on these investments.

Our

RADR® platform is rapidly emerging as a robust and scalable platform for targeted cancer therapy development. Through

the use of A.I., machine learning, and multi-agentic research systems, RADR® is designed to quickly identify and guide

the development of compounds that we can develop as potential oncology agents through either a process of drug rescue, drug repositioning

or de-novo development. RADR® is being developed through an accumulation and curation of genomic data, biomarker data,

chemical structural data, and detailed ontologically indexed documents that are directly relevant to the measurement and classification

of drug-tumor interaction, clinical datapoints related to patient response and patient stratification, and de-novo drug design.

Predicting

optimal drug responses in cancer patients requires the identification and validation of predictive biomarkers. Our RADR® platform

seeks to identify biomarkers to assist in selecting patients who have the highest likelihood to respond to our drug candidates. For example,

the targeted indications for our drug candidate LP-184 were selected in part because they are known to highly express the protein coding

gene PTGR1. Our clinical development plans for LP-184 are intended to provide additional information regarding biomarkers related to

LP-184’s molecular and cellular targets. This method of using and validating targeted biomarkers during development and then using

these biomarkers during future clinical trials can lead to shortening of the development timeline and compression of costs associated

with oncology drug development.

Similarly,

we believe LP-300 targets molecular pathways that are more common in never smokers than in other groups and also targets kinases involved

in key signaling pathways involving enzymes critical for DNA synthesis and repair, such as Excision Repair Cross-Complementation Group

1 (ERCC1), Ribonucleotide Reductase 1 (RNR1), Ribonucleotide Reductase 2 (RNR2), as well as enzymes and proteins important in regulating

cell redox status, such as Thioredoxin (TRX), Peroxiredoxin (PRX), Glutaredoxin (GRX), and Protein Disulfide Isomerase (PDI).

Our

RADR® Platform

The

human genome consists of 19,000 to 20,000 protein coding genes. One input record derived from available data bases and analyzed by our

RADR® platform consists of datapoints (expression values) from approximately 20,000 genes, another input record type is

drug sensitivity data (IC20, IC50), and other sets include key clinical parameters from HIPAA compliant patient data and clinical histories.

Our RADR® platform uses a data-driven gene feature selection methodology that is a combination of biology, informatics,

and statistics – computational biology. The architecture, tools and software of our platform are depicted in the figures below.

We

developed our platform using primarily open-source supervised algorithms such as Neural Networks, Support Vector Machine, Random Forest,

K-Nearest Neighbors, Logistic Regression and Penalized Multivariate Regression. Each algorithm is trained with labeled input data to

predict characteristics such as drug sensitivity (regressor models), stratify patient response as responder or non-responder (classifier

models), or drug behaviors such as ability to cross the blood-brain barrier (ensemble classification). Model tuning and optimization

is then performed using a hyperparameter search algorithm in order to produce the predicted lowest cross validation error. The models

are then evaluated using traditional performance metrics such as accuracy, area under the curve, sensitivity, specificity, precision,

root mean square error and mean absolute error calculations.

A

feature reduction algorithm is used to reduce the number of genes under analysis to a biomarker gene panel of less than approximately

50 genes. This set of genes is intended to carry the highest coefficient to predict drug sensitivity and the highest variable importance

in classifying a responder from a non-responder. Genes that do not help in predicting the output variable are eliminated to allow for

better prediction generalization and understand mechanisms based on key genes only.

Our

RADR®Platform Workflow

Our

RADR® platform’s proprietary workflow involves preliminary statistical analysis on approximately 18,000 features

typically from whole transcriptomic datasets and then reducing the set to approximately 2,000 features. This is followed by gene filtering

via biological and statistical methodologies yielding approximately 200 significant genes. The platform currently contains multiple feature

selection methods and multiple machine learning methods to analyze the drug and omics data, in order to fine tune the model and get better

and improved prediction accuracy. Feature selection ensures that genes that do not contribute to response prediction are excluded from

the output dataset. The prediction component subsequently applies an A.I.-driven reduction algorithm to the previously filtered genes

generating a targeted set of typically less than 50 candidate biomarkers predictive of response to a particular molecule. The figure

below illustrates RADR®’s workflow.

A

distinct and unique benefit of the RADR® platform is its ability to integrate biological knowledge and data-driven feature

selection to generate hypothesis-free biomarker signatures. This can then aid in identifying novel targets for predictive screening and

drug development.

Our

RADR® platform is enabled through access to, and analysis of, a number of key datasets: (i) publicly available databases,

(ii) data from commercial clinical studies and trials and (iii) our proprietary data generated from ex vivo 3D tumor models specific

to drug-tumor interactions. We incorporate automated supervised machine learning strategies along with big data analytics, statistics

and systems biology to facilitate identification of new correlations of genetic biomarkers with drug activity.

The

value of the platform architecture is derived from its validation through the analysis of over 200 billion oncology-specific clinical

and preclinical data points, more than 154 drug-cancer interactions, thousands of drug classes, data covering more than 135 cancer subtypes,

and over 130,000 patient records from 16 databases, one of which is our internal database. RADR® leverages this data and

over 200+ advanced ML algorithms to power its drug discovery and development modules. Our target objectives for additional data growth

efforts of the RADR® platform include a focus on drug sensitivity data, combination treatment outcome data, biomarker

data in rare cancers, and on emerging synthetic lethal targets that are aimed at accelerating the development of new therapies. Additionally,

the RADR® platform’s generative A.I. capabilities, focusing on molecular optimization and automated feature extraction

to improve understanding and prediction of molecular dynamics, safety, and drug-drug interactions are planned to increase in functionality

and scope for both small molecule development and for ADC development, analytics and characterization.

We

use cancer cell line gene expression profiles and drug sensitivity data (IC50) as one of the RADR® platform’s input

types. In a population of 10 case studies our platform was able to distinguish responders from non-responders with an average historical

accuracy of over 80%. We have also used our platform to generate genetic signatures that we believe to have applicability for the majority

of FDA approved drug-tumor indications. External validation, through retrospective data analysis, of patient datasets from 10 independent

clinical studies achieved an average response prediction accuracy greater than 80%, and internal analysis of 120 drug-tumor interactions

in cell lines achieved an accuracy of greater than 85%. The figure below illustrates examples of RADR®’s algorithms

and how they can be used.

We

have developed our platform in a cloud environment that efficiently uses parallel processing to analyze patient stratification and biomarker

selection. Best software engineering practices are followed while designing and developing our platform’s architecture. In order

to track modifications in the software, a version control system is in place. We use a software release process, including a rigorous

regression testing process, to ensure functions and programs are working as designed.

Our

platform uses a simple user input and GUI based AI architecture that can be used in many pharmaceutical research areas such as biomarker

identification, patient stratification, drug rescue and reposition by bioinformaticians, clinicians and trained wet-lab scientists.

In

late 2021, the Code Ocean Platform, a secure cloud-based computing environment manager, was integrated into RADR®. The

Code Ocean environment has upgraded RADR®’s data organization, synchronization, scalability and accessibility. These

architecture changes have enhanced the reproducibility of RADR® aided insights and analysis and created an environment

that improves the ability to collaborate and share insights within Lantern and with Lantern’s collaborators.

In

2025 and early 2026, RADR introduced two AI services intended to serve the needs of clinical researchers and biomedical scientists by

predicting which drugs can cross the blood brain barrier, and a rapid research “co-scientist” collaborator for research in

rare cancers and drug development. Both of these services provide an initial free introductory experience, with an opportunity for further

use through collaborations.

In 2026 we also plan to introduce our proprietary artificial intelligence (AI) platforms and related technologies

as a potential source of revenue. This initiative is intended to leverage our existing AI infrastructure, technologies and drug development

expertise to create new opportunities in precision oncology and translational research support. We expect to evaluate multiple partnership

and commercialization models as our platforms advance and approach market readiness through both subscriptions, access and services for

biopharma companies, researchers, drug developers, and other users.

PredictBBB

web application

PredictBBB

allows the prediction of which small molecules will cross the blood-brain barrier. It is one of many drug development modules that

we have created through the use of RADR®. With

94% accuracy in predicting which small molecules will cross the blood-brain barrier, users only need to provide a SMILES string

which represents the chemical’s structure in order to obtain a prediction. A drug name search bar provides users an easy

method to look up SMILES strings from PubChem and click to use it as the input for predictions, or users can provide their own

SMILES string for proprietary compounds. The model is based on over 4,000 molecular characteristics derived from the chemical

structure which the web application automatically generates and validates for the user input compounds, then returns a prediction

and report in less than approximately 1 minute.

withZeta.AI Rare Cancer Research Platform

withZeta

is a novel generative AI platform built to empower researchers and clinicians to dramatically improve the quality and reduce the time

of rare cancer research, therapeutic development, drug repurposing, biomarker targeting, and clinical trial design.

Built

securely in the AWS Cloud with a serverless, Lambda-based architecture and state-of-the-art large language model (LLM) inference, withZeta integrates an indexed

ontology of 438 rare cancer types, over 559,000 clinical trial records, over 204,000 published papers pertaining to rare cancers, and

532 FDA-approved oncology drug records. In conjunction with these curated databases, Zeta also utilizes a suite of tools including open-weight

specialized language models for de-novo chemical design, integration of predictBBB, molecular feature descriptions, and a collection

of targeted external resources which seamlessly integrate into a unified chat interface through task-aligned system prompts.

Actuate

Therapeutics Collaboration Utilizing RADR Platform

In

May 2021, we entered into a Collaboration Agreement with Actuate Therapeutics, Inc. (“Actuate”), a clinical stage private

biopharmaceutical company focused on the development of compounds for use in the treatment of cancer, and inflammatory diseases leading

to fibrosis. Pursuant to the agreement, we collaborated on utilization of our RADR® platform to develop novel biomarker

derived signatures for use with one of Actuate’s product candidates. As part of the collaboration, we received shares of Actuate

stock subject to meeting certain conditions of the collaboration, as well as the potential to receive additional Actuate stock if results

from the collaboration are utilized in future development efforts.

TTC

Oncology Collaboration to Expand the Clinical Development of Drug Candidate TTC-352

In

February 2023, we entered into a Collaboration Agreement with TTC Oncology (“TTC”). The collaboration focused on using RADR®

to accelerate and sharpen the drug development of TTC’s Phase 2 ready drug candidate TTC-352. TTC-352, is a novel, first-

and best-in-class selective human estrogen receptor (ER) partial agonist (ShERPA) for the treatment of patients with metastatic ER+ breast

cancer. TTC-352 was evaluated in a Phase 1 accelerated dose escalation study for hormone receptor positive metastatic breast cancer,

and it showed early anti-tumor activity signals in heavily pretreated hormone refractory patients. The aims of the collaboration were

to 1) identify biomarker or gene signatures to power potential patient selection for a planned TTC-352 Phase 2 clinical trial, 2) further

characterize TTC-352’s mechanism of action, and 3) discover additional treatment indications for TTC-352.

Oregon

Therapeutics Collaboration to Optimize Precision Development of Drug Candidate XCE853

In

mid-2024, we entered into a strategic A.I.-driven collaboration with French biotechnology company, Oregon Therapeutics, to optimize the

development of its first-in-class protein disulfide isomerase (PDI) inhibitor drug candidate XCE853 in novel and targeted cancer indications.

As part of the collaboration, we leveraged our proprietary RADR® A.I. platform to uncover biomarkers and anticancer-associated

signatures of XCE853 across solid tumors aimed at assisting in precision development. Oregon Therapeutics has focused on developing XCE853

in various cancer indications, including drug-resistant ovarian and pancreatic cancer, certain hematological cancers and several pediatric

cancers including CNS cancers. The objectives of the collaboration included a focus on 1) uncovering biomarkers and efficacy-associated

gene signatures to guide in the eventual stratification and selection of patients for future clinical trials, 2) efforts to identify

tumor-based response and resistance mechanisms to XCE853 and strategies to overcome treatment resistance, and 3) identification of opportunities

to expand the use of XCE853 in additional therapeutic cancer indications for XCE853.

Our

Strategy

Our

mission is to bring the right cancer drugs to the right patients by transforming the drug development process through the use of artificial

intelligence and data-driven development approaches. Our proprietary A.I.-enabled, and precision oncology approach, which focuses on

developing our own pipeline of compounds by rescuing drug candidates that have previously failed and developing new compounds that are

targeted to specific biological activity and genomic pathways, has the potential, we believe, to bring drugs to market faster, with lower

costs, and with reduced risk, thereby enabling a change in the cost and availability of precision cancer therapy. The strength of this

approach is demonstrated by our track record of advancing newly developed drug programs from initial AI insights to first-in-human clinical

trials in 2–3 years and at approximately $1.0 – $2.5 million per program. We work with leading research laboratories, translational

medicine and cancer centers to develop our studies and clinical trials for our portfolio, and actively update and improve our RADR®

platform to incorporate additional biomarker data, patient outcome data, cancer drug efficacy studies and computational models that relate

to oncology drug development and prediction of patient response. Our RADR® platform has grown to over 200 billion oncology-focused

data points and a library of 200+ advanced machine learning algorithms, and we are actively expanding its capabilities through new modules

focused on ADC development, combination regimen prediction, and blood-brain barrier permeability assessment.

As

part of our strategy, we plan to:

LP-300

General

Overview

We

are currently advancing LP-300 in a Phase 2 clinical trial (the “HARMONICTM Study”) of LP-300 in combination with carboplatin

and pemetrexed in never smoker patients with relapsed advanced primary adenocarcinoma of the lung after treatment with tyrosine kinase

inhibitors (TKIs).

LP-300

is a cysteine-modifying molecular entity that works to modulate multiple cellular pathways simultaneously and is a potential combination

agent for targeted indications in NSCLC. LP-300 is a small molecule (molecular weight 326.4 Da) that was in-licensed from BioNumerik

Pharmaceuticals, Inc. in May 2016, and subsequently acquired by us in 2018. We are focused on repositioning LP-300 as a potential combination

therapy for never smoker NSCLC patients with histologically defined adenocarcinoma. Prior clinical trials conducted by BioNumerik for

LP-300 did not meet their primary clinical endpoints, and at least one or more future clinical trials that meet their pre-specified primary

endpoints with statistical significance will be required before we can obtain a regulatory marketing approval, if any, to commercialize

LP-300. Safety and efficacy determinations are solely within the authority of the FDA in the U.S. or other regulatory agencies in other

jurisdictions. Currently there is no approved therapy specifically for the growing indication of never-smokers with NSCLC, and female

never smokers appear to be uniquely responsive to LP-300. With both chemosensitizing and chemoprotective activity, LP-300 has potential

as a combination agent or adjuvant in front line, second line or salvage therapy in newly diagnosed, relapsed, metastatic or advanced

NSCLC for overall survival enhancement and toxicity alleviation from primary chemotherapy or standard of care. We are currently in the

early stages of defining a specific biomarker signature that correlates with heightened sensitivity to LP-300. We believe that this signature

may help accelerate the clinical development of LP-300 and has the potential to guide patient selection for targeted clinical trials.

Prior

clinical trials conducted by BioNumerik for LP-300 did not meet their primary clinical endpoints and at least one or more future clinical

trials that meet their pre-specified primary endpoints with statistical significance will be required before we can obtain a regulatory

marketing approval, if any, to commercialize LP-300. Prior clinical trial observations are not necessarily predictive of the outcome

of any future clinical trials we may conduct.

LP-300

has been administered in multiple clinical trials to more than 1,000 subjects and has been generally well-tolerated. Retrospective analyses

of the results of a multi-country phase 3 lung cancer trial (study ID DMS32212R) in subgroups of adenocarcinoma patients receiving LP-300,

paclitaxel and cisplatin demonstrated substantial improvement in overall survival, particularly among female never smokers, where a 13.6

month improvement in overall survival (p-value 0.0167, hazard ratio 0.367) in favor of LP-300 was observed, as compared to placebo in

the subgroup of paclitaxel/cisplatin-treated patients. Similar retrospective findings of increased overall survival in the subgroup of

LP-300/paclitaxel/cisplatin treated female Asian patients with adenocarcinoma of the lung were observed in a randomized, double-blind,

placebo-controlled trial in Japan. Prior historical clinical trial observations are not necessarily predictive of the outcome of future

trials. No assurances can be given that we will be successful in obtaining marketing approval for LP-300. The chemical structure of LP-300

is depicted below.

LP-300

Chemical Structure

The

Ongoing HARMONICTM Study

We

are conducting a Phase 2 clinical trial (the “HARMONICTM Study”) of LP-300 in combination with carboplatin and pemetrexed

in never smoker patients with relapsed advanced primary adenocarcinoma of the lung after treatment with tyrosine kinase inhibitors. Our

purpose in conducting the study is to determine the potential clinical advantages and benefits for this drug combination in the study-defined

patient population. As of March 17, 2026, we have 4 clinical trial sites in the US, 5 clinical trial sites in Japan, and 5 clinical trial

sites in Taiwan. Enrollment of patients on the HarmonicTM Study in the U.S. has been challenging, and we have implemented a strategy

of increasing enrollment by expanding the study to East Asian countries where approximately 30-35+% of all lung cancer cases occur in

never-smokers with NSCLC.

The

HarmonicTM Study is designed as a multicenter, open label, Phase 2 trial with planned total enrollment of approximately 90 patients.

Patients who are never smokers with lung adenocarcinoma and have relapsed after prior treatment with tyrosine kinase inhibitors will

be eligible for enrollment. Patients who are former smokers but carry actionable genomic alteration(s) may also be eligible. Following

completion of a seven-patient safety lead-in phase in which patients received the triplet regimen of carboplatin, pemetrexed, and LP-300,

the trial advanced to the randomization stage, which consists of enrolling patients in a 2:1 allocation ratio to one of two arms: Arm

A (consisting of carboplatin, pemetrexed, and LP-300) or Arm B (consisting of carboplatin and pemetrexed). As of March 17, 2026, 24 and

10 patients have received treatment in Arm A and Arm B, respectively, of the randomization stage of the study.

The

primary objective of this study is to determine progression-free survival and overall survival in the study-defined patient population

when co-administered LP-300 with combination chemotherapy (carboplatin and pemetrexed) versus carboplatin and pemetrexed alone. The secondary

objectives of the study are to evaluate tumor response measured by objective response rate, duration of objective response, and clinical

benefit rate. We will also determine any associations between the efficacy endpoints and patient biomarkers (e.g., circulating tumor

DNA and tumor genome characteristics) as an exploratory objective. Other exploratory objectives for the study may include evaluating

quality of life in all patients and performance of patients based on the type, duration, and number of tyrosine kinase inhibitors received.

Summarized

below are some key findings from the ongoing Phase 2 clinical trial as of March 17, 2026:

● No treatment-related serious adverse events reported.

● No suspected unexpected serious adverse reactions.

In

March 2026, we submitted a Type C meeting package to the U.S. Food and Drug Administration (FDA) regarding the ongoing Phase 2

HARMONIC study. The meeting, currently scheduled for mid May 2026, seeks FDA feedback and concurrence on proposed protocol

amendments to the HARMONIC study. The proposed amendments include: (i) focusing future enrollment to patients with EGFR exon 21

L858R mutation (a subtype of tyrosine kinase mutations); (ii) increasing the maximum number of LP-300 treatment cycles from six to

eight; and (iii) converting the current randomized study design to a Phase 2 single-arm Simon two-stage study by discontinuing

enrollment into the control arm. We feel that the proposed amendments are supported by a preliminary analysis of study data

suggesting that patients with the EGFR exon 21 L858R mutation may derive greater clinical benefit from the LP-300 triplet regimen;

the evolution of the treatment landscape for TKI-refractory NSCLC that has made continued randomization to the control arm

increasingly challenging; and historical safety data indicating that up to eight cycles of LP-300 at the current dose level did not

alter the established safety profile of the drug. There can be no assurance that the FDA will concur with the proposed amendments,

and any changes to the study protocol will be subject to FDA review and clearance, during and after the Type C meeting.

Key

Findings from Prior LP-300 Clinical Trials

Summarized

below are some key findings from LP-300’s prior clinical trials:

Background-Scope

of Prior Phase 3 NSCLC Adenocarcinoma Trial (LP-300)

LP-300

was studied in a randomized, multi-center (trial locations in four US states and five European countries), double-blind and placebo-controlled

Phase 3 trial from 2010 to 2013 in patients with adenocarcinoma of the lung (the “Phase 3 NSCLC adenocarcinoma trial”). The

aim of the trial was to determine whether LP-300, combined with a standard combination of chemotherapy drugs, would increase survival

in patients with advanced NSCLC adenocarcinoma. The secondary aim of the trial was to determine if the chemoprotective properties of

LP-300 were effective in preventing or reducing common side-effects of cancer treatment, including kidney damage, anemia, nausea and

vomiting that can occur with these drug combinations. The trial enrolled NSCLC patients with newly diagnosed or recurrent advanced (stage

IIIB/IV) primary adenocarcinoma of the lung. Patients with confirmed histopathological diagnosis of inoperable and measurable advanced

primary adenocarcinoma (including bronchioalveolar cell carcinoma) of the lung, and no prior systemic treatment for NSCLC including chemotherapy,

immunotherapy, hormonal therapy, targeted therapies or investigational drugs, were included in the trial. Overall survival was the primary

outcome measure. Patients in the control arm received standard of care (cisplatin and either paclitaxel or docetaxel) plus placebo, whereas

patients in the treatment arm received standard of care (cisplatin and either paclitaxel or docetaxel) plus LP-300. The primary results

of the trial for patients receiving cisplatin and paclitaxel are outlined in the table below. While the overall results of the Phase

3 NSCLC adenocarcinoma trial did not meet the specified endpoint of the trial in increasing overall survival in all patients, when the

data were retrospectively separated by gender and smoking status, the trial data demonstrated that all never smokers, especially female

never smokers, saw increased survival with LP-300 combination treatment with paclitaxel and cisplatin. Furthermore, the LP-300 group

in the phase 3 NSCLC adenocarcinoma trial exhibited well-tolerated advantages relating to the potential to protect against chemotherapy-induced

nephrotoxicity, neuropathy and nausea along with reduced anemia.

The

figure below depicts the survival curves for cisplatin/paclitaxel subgroups for the Phase 3 NSCLC adenocarcinoma trial that ended in

2013, as summarized. The Kaplan Meier curves maintain consistent separation between treatment arms for the never smokers, females, and

female never smokers.

Rationale

Behind LP-300 Rescue and Repositioning Efforts

Based

on the results from the prior Phase 3 NSCL adenocarcinoma trial, we launched the HARMONICTM LP-300 Phase 2 clinical trial to target

the subpopulation of never smokers with adenocarcinoma that saw strong benefit in the previous Phase 3 trial. Although the incidence

of never-smokers with NSCLC is rising currently there is no approved therapy specifically for the growing indication of never-smokers

with NSCLC. Preclinical observations support that LP-300 preferentially modulates ALK and EGFR, two commonly mutated genes in non-smokers

with adenocarcinoma. Based on the findings from the previous Phase 3 NSCL adenocarcinoma trial, it is possible that the benefits of combining

LP-300 with standard of care chemotherapy could be further improved by identifying additional molecular biomarkers in patients who respond

well to LP-300 combination treatment. We continue to seek additional opportunities for LP-300. Some of our considerations include a never

smoker population with a specific genetic signature that correlates to increased LP-300 sensitivity.

Disease

Background and Opportunity

Lung

cancer is the second most prevalent cancer globally, and it accounts for the highest level of cancer-related deaths worldwide. Lung cancer

accounts for approximately 12% of all new cancer diagnoses, but 21% of all cancer deaths in the US. Lung cancer kills more people annually

than cancers of the breast, prostate, colon, liver, kidney, pancreatic, and melanoma combined. The American Cancer Society’s estimates

for lung cancer in the US for 2026 are:

The

most common type of lung cancer is called non-small cell lung cancer (“NSCLC”), which represents about 80% to 85% of all

lung cancer.

Lung

adenocarcinoma, a histological subtype of NSCLC that originates within the glands that line the lung, is the most common subtype of lung

cancer in the world inflicting approximately 50% to 65% of non-Asians and approximately 70% to 85% of Asians diagnosed with lung cancer.

According to LUNGevity Foundation, the National Institutes of Health and other published literature, 60% to 65% of all new lung cancer

diagnoses are among people who are former smokers or have never smoked, while 10-15% of new lung cancer cases are among never-smokers.

Over