Technology

Technology 2018-11-12T09:13:21+00:00

RADR or Response Algorithm for Drug Positioning &  Rescue is Lantern's proprietary integrated data analytics, experimental biology and machine learning based platform for patient genetic profiling for robust drug response prediction.

RADR is comprised of 4 key components

RADR-NET

  • Patient data from clinical trials are analyzed to determine drug responses, clinical efficacy, and safety and closely matching tumor samples are obtained from clinical network

RADR-3DLab

  • Lab-based genomic and drug sensitivity analysis of patient samples and diseasespecific genetically modified tumor models

RADR-AI

  • Our Al-based machine learning approach combines of three automated modules that work sequentially to derive drug and tumor-specific complex biomarker panels. These three main modules include: data pre-processing, feature selection, and prediction.(a) Data pre-processing includes data cleaning, transformation and normalization without compromising the original quality of data.
    (b) Feature selection, RADR-AI performs proprietary gene filtering via biological, statistical and machine learning-based methods. Not all genes have equivalent relevance for response prediction and this method ensures that genes that do not contribute to outcome prediction are excluded from the output.
    (c) Prediction AI-driven program reduces the initial number of genes (approximately 500) to a more manageable number, typically less than 50 potential predictive biomarkers.

TRIALS and STRATIFICATION

  • Biomarker panels are derived to select true responders for recruitment into clinical trials. RADR technology can derive robust, reliable, reproducible biomarker panels to select true responders for recruitment into clinical trials improving the success rate of clinical approval which is compatible with FDA guidelines.

Validation

The current RADR platform version 3.0 has been developed and tested on more than 120 drug-cancer interactions. RADR has been validated using several clinical trial datasets present in the GEO database, with analysis of over 400 patient records.  Analyses of patient data achieved an accuracy rate greater than 80% during external validation and cell line based 120 drug-tumor interactions achieved an accuracy rate greater than 85% during internal validation.

Loading...

Download Our White Paper

To Download our RADR™ PLATFROM WHITE PAPER please fill in the from below.

Recent Publications

Title: Response Algorithm for Drug positioning and Rescue (RADR™): Lantern Pharma's Artificial Intelligence based integrative machine learning approach for drug positioning and rescue

Conference: American Association for Cancer Research (AACR) Special Conference on Convergence: Artificial Intelligence, Big Data, and Prediction in Cancer, taking place October 14 – 17, 2018 in Newport, RI.

Presenter/Author: Umesh Kathad, MS
Session: Session A, Poster Board Number: A07

Session date and time: Session A, Monday, Oct. 15 from 5:30-7:30 p.m.

  • RADR™ uses transcriptomic, drug sensitivity and systems biology inputs and generates gene expression-based responder/ non-responder profiles for specific tumor indications.
  • RADR™ comprises three main modules: data pre-processing, feature selection, and response prediction.
  • RADR™ has demonstrated a response prediction accuracy greater than 80% during clinical validation.

Read More

Title: External validation of Lantern Pharma's Response Algorithm for Drug positioning and Rescue (RADR™) using Paclitaxel clinical data

Conference: American Association for Cancer Research (AACR) Special Conference on Convergence: Artificial Intelligence, Big Data, and Prediction in Cancer, taking place October 14 – 17, 2018 in Newport, RI.

Presenter/Author: Umesh Kathad, MS
Session: Session B, Poster Board Number: B49

Session date and time: Tuesday, Oct. 16 from 5:15-7:15 p.m.

  • Retrospective analyses were performed as part of RADR™ validation using 4 independent datasets of breast cancer patients treated with Paclitaxel either as a monotherapy or as a combination.
  • In one example of a Paclitaxel combination trial with breast cancer patients, from only 109 patient records, RADR generated a signature of 45 genes, whose expression level was predictive of Paclitaxel combination treatment response with responder prediction accuracy of 85%.
  • Application of the RADR™ program to this Paclitaxel trial could have potentially reduced the number of patients in the treatment arm from 109 unselected patients to 50 biomarker-selected patients to produce the same number of responders.

Read More

Title: Novel method for predicting sensitivity to Lantern Pharma's pipeline drug candidate using the Response Algorithm for Drug positioning and Rescue (RADR™)

Conference: American Association for Cancer Research (AACR) Special Conference on Convergence: Artificial Intelligence, Big Data, and Prediction in Cancer, taking place October 14 – 17, 2018 in Newport, RI.

Presenter/Author: Aditya Kulkarni, PhD
Session:
 Session B, Poster Board Number: B51

Session date and time: Tuesday, Oct. 16 from 5:15-7:15 p.m.

  • RADR™ emphasizes the integration of biological knowledge, data-driven feature selection, and robust Al algorithms to achieve hypothesis-free biomarker identification.
  • As part of RADR™ drug model building and development, we used a dataset showing preclinical efficacy of our pipeline oncology candidate LP-184.
  •  Using NCI-60 cell line gene expression profiles covering >18,000 transcripts, and proprietary data on LP-184 sensitivity, RADR™ identified a panel of 10 genes whose expression levels are predictive of response to LP-184 with 100% accuracy.
  • Genes from this panel of 10 candidate biomarkers were found to be functionally involved in LP-184-specific induction of bioactivation and are in agreement with the known mechanism of action of LP-184.

Read More