Reading X-Rays
with Intelligent Eyes
How a low-code imaging platform built on Zoho Creator — powered by Google Gemini's multimodal AI — is giving clinics and hospitals the ability to analyse medical scans through plain-language prompts.
A radiologist stares at a chest X-ray at 11 PM, fatigued after a twelve-hour shift. A junior doctor in a rural clinic has a CT scan but no specialist available for three days. A hospital administrator needs to retrieve scan notes for a patient who was treated six months ago. Each of these situations carries real risk — and each one can be meaningfully improved by the right technology.
The Challenge with Medical Imaging Today
Medical imaging — X-rays, CT scans, MRIs, and ultrasounds — produces an enormous volume of data every single day in any active clinical setting. The challenge is not generating the images. The challenge is analysing them accurately, consistently, and quickly enough to actually serve patient care.
Traditional workflows involve a radiologist reviewing each image, dictating findings, a transcriptionist creating a report, and that report reaching the treating physician hours or even days later. In busy hospitals, backlogs are common. In smaller clinics, specialist access is limited. And everywhere, the accuracy of documentation depends heavily on how tired or rushed the reviewing clinician is.
What if a clinic could upload a patient's X-ray, type a question, and receive a structured medical analysis within seconds? That is precisely what becomes possible when Zoho Creator's low-code application platform is connected to Google Gemini's multimodal AI API.
"AI does not replace the clinician's judgement — it gives them a second set of eyes that never gets tired, never gets rushed, and never misses a pattern it has been trained to see."
What We Built in Zoho Creator
Zoho Creator is a low-code application development platform that allows clinics and healthcare organisations to build fully functional, secure, role-based web and mobile applications — without requiring a team of software engineers. Think of it as a medical records workbench that your practice manager can build and adapt, rather than waiting months for IT to develop something.
For this solution, we created two connected forms inside a single Zoho Creator application:
Form 1 — Patient Image Upload
This form acts as the intake point for all medical imaging files. Each record captures:
Image Upload Field
Accepts X-ray, CT scan, MRI, or ultrasound images in common formats (JPEG, PNG, DICOM-exported).
Patient Reference
Patient ID or name, date of scan, and the referring doctor — creating a traceable, auditable record.
Scan Type
A dropdown indicating whether the image is a chest X-ray, abdominal scan, skeletal image, or neurological scan.
Clinical Notes
A free-text field for the attending staff to record any initial observations or clinical context.
Here is what the upload form looks like inside the application:
Click to upload or drag and drop your scan here
Supports JPG, PNG · Max 20 MB
Form 2 — The AI Analysis Prompt
The second form is deliberately simple. It presents the clinician or radiologist with two inputs: a lookup to select the patient's uploaded image, and a plain-text area labelled "What do you want to know about this scan?"
Once submitted, Zoho Creator's backend workflow takes over — retrieving the image, building a structured prompt, and dispatching everything to Google Gemini's multimodal API. The AI's analysis is returned and displayed directly in the application, and optionally saved as a permanent record against the patient file.
How the Integration Works — Simply Explained
Google Gemini is a multimodal AI — meaning it can process and understand both text and images simultaneously. Unlike older AI models that only read text, Gemini can literally look at a medical scan and describe what it sees, identify structural anomalies, and respond to specific clinical questions about the image. This is what makes the integration so powerful for healthcare.
Here is the end-to-end process in plain steps:
Image is uploaded and stored
The patient's scan is uploaded via the Zoho Creator form. The image file is stored securely within Zoho's cloud infrastructure and referenced by a unique record ID.
Clinician types a natural language prompt
Using the second form, the doctor or radiologist selects the patient record and types their question — no special syntax or technical knowledge required.
Zoho Creator's workflow assembles the API call
A Deluge script fetches the image as a base64-encoded file, combines it with the prompt text, and formats the full request according to Gemini's API specification.
Gemini analyses the image and the question together
The multimodal model processes both simultaneously — it sees the image and understands the clinical question, responding with findings, anomalies, and observations relevant to what was asked.
The analysis appears instantly in Zoho Creator
The AI's structured response is displayed on screen and saved as a linked record within the patient file — creating a searchable, auditable history of all AI-assisted analyses.
What Clinicians Can Actually Ask
The prompt interface is intentionally open-ended. Any clinical question can be submitted. Here are examples of the kinds of prompts your team might use in daily practice:
"Describe any abnormalities visible in this chest X-ray. Are there signs of consolidation, pleural effusion, or cardiomegaly?"
"Identify any fractures or stress injuries in this skeletal scan, and indicate their likely severity and location."
"Are there any visible masses, organ enlargements, or vascular abnormalities in this abdominal CT image?"
"Compare this scan with the notes on record and indicate whether the condition appears to have improved, worsened, or remained stable."
And here is an example of what a Gemini response might look like when returned inside the Zoho Creator application:
Overall Assessment: The chest X-ray demonstrates findings consistent with right lower lobe pneumonia with early consolidation. Cardiac silhouette appears within normal limits. No pneumothorax identified.
This structured output — generated in seconds — gives the treating physician a clear starting point. It does not replace the radiologist's formal report, but it provides immediate clinical direction, especially valuable in urgent or after-hours scenarios.
The Advantages for Clinics and Healthcare Organisations
Both platforms bring complementary strengths that, when combined, create a genuinely practical clinical tool:
| Advantage | What It Means in Practice |
|---|---|
| Faster Preliminary Analysis | Clinicians receive a structured AI analysis of any scan within seconds of uploading — reducing reliance on immediate specialist availability for initial triage. |
| Reduced Diagnostic Fatigue | AI provides a consistent second opinion that does not degrade with fatigue, workload, or time of day — supporting radiologists on high-volume days. |
| Instant Information Retrieval | Any past scan can be pulled up and queried in plain language — no complex PACS navigation, no waiting for reports to be located and shared. |
| Low-Code Flexibility | Zoho Creator allows clinic administrators to adapt forms, add new scan types, or adjust workflows without engaging a development team — keeping costs low. |
| Searchable Audit Trail | Every prompt and AI response is saved as a linked record — providing a transparent, searchable history of all AI-assisted decisions for medico-legal purposes. |
| Accessibility for Smaller Clinics | A rural GP clinic with limited specialist access can now get a structured preliminary read on an X-ray and make a more informed referral decision immediately. |
| Scalable Infrastructure | The same application can manage 20 scans a month for a small clinic or 2,000 a month for a hospital network — with no architectural changes needed. |
Important clinical note: AI-generated image analysis is intended as a decision-support tool, not a diagnostic substitute. All AI findings should be reviewed and confirmed by a qualified clinician before informing patient care. Organisations should comply with applicable medical device regulations and data privacy laws (including HIPAA, DPDP, or GDPR as relevant) when processing patient imaging data through any external API.
Beyond X-Rays — Where Else This Applies
The imaging analysis pattern built here is directly extensible to other clinical and administrative use cases within the same Zoho Creator application:
Neurological Imaging
Query MRI brain scans for visible lesions, white matter changes, or structural asymmetries relevant to neurology referrals.
Orthopaedic Scans
Ask about joint space narrowing, bone density concerns, or post-surgical hardware positioning in skeletal images.
Cardiac Screening
Analyse echocardiogram images for valve abnormalities, wall motion defects, or chamber enlargement indicators.
Discharge Summaries
Use AI to automatically generate structured scan summaries for discharge letters — pulling from the analysis history on file.
Each of these extensions requires only a new scan type in the dropdown field and, optionally, a different system prompt to guide Gemini's clinical framing. The infrastructure you build once serves every specialty you choose to extend it into.
Intelligence Where It Matters Most
Healthcare has always been about human expertise, trust, and care. What AI changes is not who makes the decisions — it is how quickly and accurately those decisions can be informed.
By building a structured imaging repository in Zoho Creator and connecting it to Google Gemini's multimodal intelligence, clinics of any size can give their clinical teams a powerful, accessible, and auditable AI companion — one that works around the clock, responds to plain language, and never loses a patient record.
The technology is available today. The implementation is achievable without a large IT investment. And the potential benefit — faster analysis, fewer missed findings, more confident clinical decisions — is directly measurable in patient outcomes.
Start with one form. Upload one scan. Ask your first question. See what is possible.