A working checklist for founders setting up a new diagnostic centre or radiology unit. Eight steps, written by the team that runs the radiology backbone for 1,500+ centres across India.
TL;DR
Setting up a diagnostic centre in India takes 6-9 months and ₹40 lakh to ₹12 crore in capex depending on modality scope. The eight steps are: scope, business plan, premises, licensing, equipment, staffing, radiology infrastructure, soft-launch.
The single most common mistake first-time operators make on the radiology side is committing to fixed costs before the volume can absorb them — a full-time radiologist on day one (who you can't yet keep busy), and scanner-bundled PACS that locks you in. A pay-per-scan partnership with a radiology network (like 5C Network) gives you PACS, AI, and 400+ specialists from day one with zero capex — and you bring in-house radiologists on later, once sustained scan volume can fill their day.
India is short on radiology capacity. The country has roughly 100,000 diagnostic centres and fewer than 15,000 radiologists, and the gap is widening — scanners get cheaper and more plentiful every year, radiology training does not. For founders, this is an opening: there is genuine, durable demand for new centres in tier-2 and tier-3 cities, in suburban catchments around tier-1 metros, and inside hospital tie-ups that have outgrown their in-house imaging.
But starting a diagnostic centre is not just an equipment-buying exercise. You are setting up a small clinical operation that has to satisfy AERB, the Clinical Establishment Act, PNDT, biomedical waste rules, and corporate empanelment standards — and produce reports, not just images. The reporting layer is where most first-time operators either over-spend (full-time radiologist on day one) or under-build (no PACS, no AI, manual workflow) and where the long-term economics of the centre are actually decided.
This guide is the working checklist we wish first-time operators had. It is the eight steps we see successful centres take, in order, and the five mistakes we see most often. The radiology infrastructure piece is where 5C Network plays — Step 7 covers it honestly.
The Setup
Order matters. Skipping a step or doing one out of sequence costs months and lakhs.
Pick modalities based on catchment, not ambition.
Standalone digital X-ray + ultrasound: ₹40-60 lakh capex. Add CT: ₹3-5 crore. Add 1.5T MRI: ₹8-12 crore. Modality mix dictates floor area, power supply, and AERB shielding — decide before signing a property lease.
Model break-even, not best-case.
Most well-located CT/MRI centres in tier-2 cities break even at 60-70% utilisation in 14-18 months. Plan working capital separately (3-6 months opex). Model radiologist costs as a separate line — first-time operators almost always over-spend here.
Ground floor, AERB-compliant shielding, RCC slab for MRI.
Minimum 2,500-4,000 sq ft for multi-modality. Dedicated 3-phase power with backup. AERB site-plan approval BEFORE pouring concrete. Civil works for radiology rooms take 8-12 weeks. Get this wrong and you redo shielding mid-construction.
AERB, Clinical Establishment Act, PNDT, Pollution Board, Fire, GST.
AERB licence for CT/X-ray/mammography. State Clinical Establishment Act registration. PNDT registration is mandatory the moment you do ultrasound (sex determination is a non-negotiable). Pollution Control Board for biomedical waste. NABL accreditation is optional but unlocks corporate empanelment. Budget 4-6 months and ₹3-5 lakh.
Refurbished is OK. Vendor-locked PACS is not.
Refurbished CT/MRI from Wipro GE, Siemens Healthineers, Philips can cut capex 40-50%. Insist on DICOM-3 output. Cap AMC at 8-10% of capex annually. Refuse "free" bundled PACS — they lock you in and break AI integration later.
Technologists first. Radiologists later.
2-3 radiographers per modality (for two shifts), centre manager, 2-3 front-desk/billing staff, biomedical engineer (full-time or contract). Hire technologists 4-6 weeks before launch so they train on your scanner models. For radiologists, see Step 7.
PACS + AI + reporting. This is where 5C Network plugs in.
You need three things, ideally from one partner: a cloud-native PACS to receive and route DICOM studies, AI for triage and pre-reads so technologists know what is urgent, and on-demand radiologist reporting from a credentialed panel — 24/7, all modalities, sub-30-minute turnaround. A single in-house radiologist, however good, cannot personally cover nights, weekends, holidays, and every subspecialty (neuro, MSK, paediatric, breast) — a network of 400+ specialists can. Most successful centres at scale run both: in-house for sustained-volume modalities, network for after-hours, overflow, and subspecialty cases. Pay-per-scan with 5C Network gives you that network on day one with zero capex.
Operations excellence drives referrals. Marketing does not.
Soft-launch with 4-6 referring physicians in a 5 km radius. Track TAT, no-show rate, scan-to-report time, repeat rate from week one. Hit 100 scans/day per modality before investing in marketing. Plan a second location only after first site clears 80% utilisation.
Watch outs
Across 1,500+ centres on the network. In order of how expensive they are to undo.
A great in-house radiologist is gold once you can keep them busy. On day one — with mixed modalities and unpredictable load — you are paying ₹25-40 lakh/year for capacity you cannot use, and the radiologist is underutilised. Bad for both sides. Start with on-demand reporting from a network so coverage is 24/7 from week one, then add in-house talent once sustained volume can fill their day. The two are complementary, not exclusive.
Free PACS that comes "free" with your CT/MRI is the most expensive software you will ever own. It locks you into the vendor, breaks AI integration, and forces a painful migration when you add a second scanner brand.
Pouring concrete before AERB approval is the single most expensive mistake a new centre can make. Redoing shielding mid-construction costs ₹15-30 lakh and 6-10 weeks.
Bad positioning produces bad studies, which produce repeat scans, which produce angry patients and low referral pull-through. Train technologists on your specific scanner models for 4 weeks before going live.
A second location at 60% utilisation of the first is a recipe for two underperforming centres. Push the first site past 80% utilisation, then expand.
"The first-time operators I see succeed are the ones who decide the radiology infrastructure question on day one — not month nine when scan volume already hurts. PACS, AI, and reporting are not back-office software. They are the operating layer of the centre, and they decide whether your turnaround time is 30 minutes or 30 hours."
Co-founder & CEO, 5C Network
FAQ
Capex depends entirely on modality scope. A digital X-ray plus ultrasound centre costs ₹40-60 lakh. Adding a CT scanner pushes the budget to ₹3-5 crore. A full multi-modality centre with 1.5T MRI lands at ₹8-12 crore. On top of capex, plan for ₹15-30 lakh in licensing, civil works, and 3-6 months of working capital. The single biggest under-budgeted line item for first-time operators is radiology infrastructure: PACS, AI, and reporting workflow.
At minimum: AERB radiation safety licence (for CT, X-ray, mammography), state Clinical Establishment Act registration, PNDT Act registration (mandatory for any centre doing ultrasound, to prevent sex determination), Pollution Control Board clearance for biomedical waste, Fire department NOC, GST registration, and Medical Council registration for the centre's resident radiologist. NABL accreditation is optional but unlocks corporate and TPA empanelment and improves price realisation. The full licensing cycle takes 4-6 months and roughly ₹3-5 lakh in fees and consultancy.
Not on day one. A full-time MD radiologist in India costs ₹25-40 lakh per year, and a new centre with mixed modalities and unpredictable volume usually cannot fill their day — so you are paying for capacity you cannot use, and the radiologist is underutilised. Start with on-demand reporting from a teleradiology network so coverage is 24/7 from week one (including nights, weekends, holidays, and subspecialties like neuro, MSK, paediatric, breast). Bring an in-house radiologist on once your sustained daily volume can absorb a full day of reads in modalities they are trained to interpret — and even then, keep the network for after-hours, overflow, and subspecialty cases. Most well-run centres at scale run both. The two are complementary, not exclusive.
A new centre should run on a cloud-native, vendor-neutral PACS — not the proprietary PACS bundled with your scanner. Vendor-neutral PACS gives you DICOM-3 output, easy integration with AI tools and teleradiology partners, no per-modality licensing traps, and the freedom to swap scanners later without losing your image archive. Pair it with AI for triage and pre-reads on common pathologies (chest X-ray, head CT, knee MRI) so your technologists know what's urgent. 5C Network bundles PACS, AI, and on-demand radiologist reporting in a single pay-per-scan workflow — that's the simplest setup for a new centre.
Plan for 6-9 months from incorporation to first paid scan. Civil works and shielding take 8-12 weeks. AERB and Clinical Establishment Act licensing take 4-6 months in parallel. Equipment delivery and commissioning takes 6-10 weeks (longer for MRI). Staff hiring and training takes 4-6 weeks before going live. The critical-path items are usually AERB approval (start early) and MRI delivery (book the order before pouring the slab).
Ultrasound is almost always yes — it's the highest-volume modality in most catchment areas and has the fastest payback on capex. Mammography becomes worth adding once you cross 40+ years female footfall of 30+/day; otherwise the screening volume doesn't justify the capex. DEXA is a niche add-on that breaks even slowly — pursue it only if you have an existing orthopaedic or endocrinology referral pipeline. The right modality mix is dictated by your catchment's demographics and your referring physicians' actual ordering patterns, not by what's commercially trendy.
Yes. Most centres on the 5C network are single-site operators in tier-2 and tier-3 cities. The pay-per-scan model means there is no minimum volume commitment, no upfront fee, and no capex for PACS or AI. A single-modality centre can go live on the platform in under 72 hours. 1,500+ hospitals and diagnostic centres across 27 Indian states currently report with 5C, with a median report turnaround of under 30 minutes.
If you found this useful, these go deeper on the radiology layer.
Guide
Five mandates, three provider archetypes, and an SLA checklist for buyers.
Whitepaper
A framework for setting up AI-native radiology workflows from day one.
Calculator
Model what pay-per-scan would cost for your centre's scan volume.
We will help you scope PACS, AI, and reporting for your modality mix — and tell you honestly whether a full-time radiologist makes sense for your scan volume. Free 30-minute call, no slide deck.
