---
AI Summary: "\\- Conducted end‑to‑end AI/ML development for cardiomegaly detection\
  \ using MONAI and self‑supervised key‑slice segmentation.  \n- Transformed research\
  \ algorithm into FDA‑compliant, deployable software, handling cybersecurity and\
  \ regulatory communications.  \n- Executed clinical performance study and secured\
  \ 510(k) clearance, delivering a production‑ready AI solution."
Anonymous: false
Assignee:
- Yujan Shrestha
Last Edited Time: '2026-01-25T23:07:00+00:00'
case_study_510ks:
- 7172f211-9d74-4db2-97f3-5d32a347bc5e
client_logo:
- 18dbd5b7-a754-809a-85b7-f58c187fa681
client_name: Body Check
date: '2024-11-22'
featured: true
medical_panel: Radiology
name: 'We Did It All: R&D, Clinical Study, and 510(k) Cleared in 12 Months'
services:
- 124bd5b7-a754-8044-ba47-cac680c0e88e
- 9dc4f55a-b5cb-419f-8bae-5e11d6b11fee
- 11abd5b7-a754-8058-b035-f358222cec8c
- 299bd5b7-a754-8036-8380-d6165b0c1670
summary: 'Built and cleared an AI cardiomegaly detector end-to-end: algorithm R&D
  to 510(k) clearance in under 12 months.'
tags:
- AI/ML
- 510k
- Image Processing
testimonials:
- 141bd5b7-a754-80ce-bd20-d6b958655c69
---

## The Problem

Drs. Rothenberg and Dr. Andrew Smith at UAB identified a need to address the global cardiovascular disease (CVD) burden using AI for opportunistic screening during routine imaging. Given limited primary care access and late diagnoses, they proposed AI-based image processing to detect **cardiomegaly**, an indicator of underlying CVD. Standard care doesn\'t assess cardiomegaly due to its cumbersome measurement, making it an ideal candidate for AI. This approach, focusing on early detection via cardiothoracic ratio measurement, aims to improve health outcomes and reduce CVD progression.

<figure>
  <img src="/img/portfolio/We_Did_It_All_RD_Clinical_Study_and_510k_Cleared_in_12_Months-141bd5b7a75480f0a61bc1cb2fe0b6e7.png">
  <figcaption>
    Dr. Steven Rothenberg on the Monday edition of RSNA daily.
  </figcaption>
</figure>

## What We Did

Innolitics did the entire project end to end---everything from algorithm R&D, software engineering, regulatory clearance, and clinical study design and coordinated the study execution.

<figure>
  <img src="/img/portfolio/We_Did_It_All_RD_Clinical_Study_and_510k_Cleared_in_12_Months-141bd5b7a754805cb5aeffab47dcb454.png">
  <figcaption>
    We developed a novel self-supervised method to determine the key slice
    to segment using just 2D images. In addition to the training pipeline,
    we also designed the inference pipeline that runs quickly even on a
    small dual core 4GB instance without a GPU.
  </figcaption>
</figure>

- Trained the algorithm and performed multiple rounds of R&D at UAB
- Developed an AI/ML algorithm from data and annotations collected at UAB using tools such as MONAI
- Transformed the research-grade algorithm into an FDA-compliant, deployable solution
- Consulted the UAB research team to collect the appropriate data necessary to substantiate an 510(k) submission
- Conducted an FDA presubmission meeting to validate study design before execution
- Created and submitted the 510(k)
- Handled all FDA cybersecurity concerns
- Handled all FDA communications including hold letter response strategy and defense

## Timeline

March 6 2023: FDA Presubmission Meeting

June 28 2023: Algorithm development done

July 26 2023: Clinical Performance Study Completed

August 28 2023: 510(k) Submitted

November 3 2023: FDA Hold Letter received

February 28 2024: 510(k) Cleared

<figure>
  <img src="/img/portfolio/We_Did_It_All_RD_Clinical_Study_and_510k_Cleared_in_12_Months-141bd5b7a7548019af9bd19e691c8c3d.png">
  <figcaption>
    FDA clearance letter.
  </figcaption>
</figure>

## About Steven

Dr. Steven Rothenberg is an Assistant Professor in the Cardiopulmonary Imaging Section at the University of Alabama at Birmingham (UAB) Department of Radiology and a pioneer in radiology innovation. He holds a Bachelor\'s in Physics from the University of Maryland and a medical degree from George Washington University. During his residency in Diagnostic Radiology, Dr. Rothenberg co-founded EnvoyAI---the first radiology AI marketplace---which was later acquired by TeraRecon.

Dr. Rothenberg\'s research focuses on imaging informatics, population health screening, and non-pharmacologic pain interventions. Recognized for his entrepreneurial spirit with accolades such as the Department or Radiology Light Bulb Award and UAB Harbert Institute for Innovation and Entrepreneurship Most Prolific Inventor, Dr. Rothenberg continues to push the boundaries of innovation in radiology and healthcare technology.