A study, a colleague and a conversation, one swipe apart

When an imaging study needs a second opinion now, the slowest part is rarely the medicine. It is finding the right colleague, getting the images in front of them, and knowing they have actually seen them. For a medical-imaging communications company, Toobler built the iPhone app and the real-time backend that put a study, a colleague and a conversation one swipe apart.

Customer and context

The client served physicians who work with medical imaging, and already ran an OsiriX server — Mac software that stores and serves studies in DICOM, the standard format for medical images. Sharing a study with another physician, and knowing it had landed, still meant phone calls and separate tools.

The client’s brief was blunt about what mattered: some cases are urgent, and the sender needs to know the recipient has the details. Every action had to take the fewest possible swipes.

Toobler was the engineering team from early 2015 to mid-2016. The client provided a starting design and ran the OsiriX side; Toobler built the native iOS app and the Node.js backend behind it, and refined the design along the way.

The challenge

• Imaging data is heavy. A single study is many DICOM images, downloaded as a compressed archive. A phone has to fetch it, unpack it, organize it and display it without running out of memory.
• Images have to look right. DICOM covers many kinds of equipment, and each has its own conventions for how pixel values become an image. Getting one wrong produces a picture that looks plausible and is not.
• Urgency needs proof of delivery. Sending a study is not enough. The sender needs to see that it arrived and that it was opened.
• A conversation should not mean switching apps. Physicians needed to move from a message to a call to showing their screen, about the study in front of them.
• The data is sensitive. Patient imaging and messages between clinicians must be protected as they move between the app and the server.

The solution

Swipe-driven actions on every study. Swipe one way to set a follow-up: tomorrow, in three days, in a week, or at a chosen date and time. The app raises a notification when it is due and opens directly on that study. Swipe the other way to download it, share it with another physician, or see its details.

A DICOM viewer built for the phone. The app authenticates against the OsiriX server, downloads studies, and unpacks them on the device — resuming if the app is backgrounded mid-way. It separates each download into its studies and series, converts the images with the Imebra DICOM library, and stores them locally for fast viewing. The team spent real effort on memory: tracking down leaks, moving to a newer version of the imaging library, and reworking how images were stored so that large studies stayed usable. One fix shows the kind of detail involved: computed-radiography X-rays were coming through with black and white inverted, and the pipeline was corrected to render them the right way round.

Communication anchored to the study. A Node.js backend using Socket.IO carries one-to-one and group chat, with delivery status, history, search and audio messages. Audio and video calls run on OpenTok, with ringing, the ability to switch between chat and a call mid-conversation, and screen sharing so one physician can walk another through what they are seeing. The backend handles dropped connections and callers who are already on another call.

Delivery you can see. Messages show when they have been delivered. Shared studies carry a view status, and the server keeps a list of studies that were shared but not yet received — the client’s urgency requirement expressed as data the sender can check.

Protection in transit. Data moving between the app and the server is encrypted with AES and RSA, implemented to work across both iOS and Node.js. The app is protected by a passcode.

Operating it. Activity logs, CSV export to Dropbox, server monitoring with alerts, and an admin page for setting up hospital regions supported the product as it moved through testing.

A key workflow: sharing an urgent study

1. A physician opens a study downloaded from the OsiriX server and reviews it in the in-app viewer.
2. They swipe, choose share, and search for the colleague.
3. They add a message and send it.
4. The colleague is notified; the sender sees the message delivered, and the study’s view status changes when it is opened.
5. If it needs discussion, either side moves from chat to an audio or video call and shares their screen to talk through the images.

Results

Toobler took the iOS app from first screens through to App Store submission, with the backend, real-time communication and imaging pipeline in place, tested across multiple users on beta builds. Work on an Android version followed in 2016.

This is a build of its time, from 2015 and 2016. The engineering it shows is still what this kind of product demands: imaging data handled carefully on a constrained device, real-time communication built around the case rather than beside it, and delivery that the sender can verify.