On September 27, at ASTRO in Boston, Siemens Healthineers launched Accela, and it is an amazing machine. The company describes multi-site SBRT in as little as 60 seconds. What I saw in Boston adds three things the public pages do not list yet: an MLC with no interleaf leakage, a collimator that rotates, which opens up a lot in what the machine can do, and a design that is easier to service.

It was co-developed with seven institutions, Mayo Clinic and Memorial Sloan Kettering among them, and the people who built it have earned the attention it is getting.

Two more facts belong beside the launch. The product page reads: "510(k) pending. Not available for sale in any market." And no price and no service terms have been published. Any figure for either this fall is an estimate, and should arrive labeled as one.

Where I stand, in my own words: "We want companies to always innovate. New innovation pushes the field forward. New innovation also makes current platforms more affordable, and those platforms are not new and were the standard of care until the new shiny one. They are still amazing."

And the part that matters for this issue: "You should map your own needs and future strategy and pick the best and incorporate new technologies into your own roadmaps."

I spent thirteen years at Varian. Asked what a launch does for the customers who do not buy, this is my answer: "It does two things. It creates pressure to buy the 'latest greatest' so as to not fall behind, but it simultaneously opens the door to better negotiation positions."

The second half belongs to a later conversation, once there is a quote. This issue is about the first, because the pressure shows up months before any paper does.

The pressure usually arrives as a question: is the new one worth it. Asked in the abstract, it has no answer, and neither does the objection about price. Both start from the machine instead of from the program.

The rest of my answer: "In the case of this machine, 'more time' is only valuable if that is something the specific program needs. Think of that across the board of each feature and it becomes clear that innovation is amazing but only in the context of the specific customer needs."

And the platform that was the standard of care on September 26 did not get worse on September 27. For a program whose map points there, it is an excellent machine to buy, or to keep running longer, and the launch tends to make that easier to afford.

So take it feature by feature, the way that answer says to.

Gantry speed and dose rate. 2.5 RPM and up to 40 Gy/min shorten the time the beam is on. That is worth the most where the day waits on the beam, in stereotactic and breath-hold work. Where setup, imaging and getting the patient on and off the table set the length of the slot, it changes less.

The MLC and the collimator. A newly designed dual-layer MLC with faster, finer modulation and no leakage between leaves, and a collimator that rotates, matter most in complex plans: several targets in one course, sharp dose fall-off next to critical structures.

Imaging and adaptive. The largest HyperSight imaging panel they have built and a 15-second cone-beam CT shorten the part of every fraction that happens before the beam and make a single breath-hold practical. The same imaging carries the adaptive capability Siemens Healthineers describes, on-couch and off-couch: re-planning to the patient's anatomy that day, with the patient on the table. That is a real clinical step. It also adds work the program carries on every adapted fraction, the staff time to review and approve the new plan at the console, and the payment does not go up with it. So adaptive belongs on the map where the case mix needs it, with that staff time written beside it.

Planning. The planning software lets the team decide, patient by patient, whether a plan aims for the best possible dose or the fastest treatment. That helps a program with both kinds of patients: complex cases that need precision, and a full schedule that needs time back.

Serviceability. This is the one that touches every program, because downtime does. A day the machine is down is a day of patients moved, rescheduled or sent elsewhere, however fast the machine is when it runs. Read it against your current machine's downtime record, which your service history already holds.

For each of these, the test is what the gain turns into on your schedule: another patient, a shorter day for the team, room for a service line already on the plan, or nothing, because that part of the day was never the constraint.

Six in ten of the new administrators in the course Heather Turner and I taught at SROA in Boston said a capital request was one of the situations they were walking into. Many of those requests were built before September 27, and most will now need an updated quote, even if only for comparison. That makes this the moment to write your own page, before the new quotes arrive and set the agenda.

The people who should help write the page already work in your building: the physicians who refer and treat, your physics lead, your lead therapist, and whoever owns the five-year plan. Commissioning and QA time for any new platform goes on the page as a planning line, sized by your physics team, the same way construction time does.

If the page says a machine comes out of service in the next cycle, Heather's column below is the floor half of that plan: what the swap costs in displaced patients, how patients move between vaults during construction, and how to bring the new machine up in phases.

Read against the page, the current platform will often win this cycle, and that is a good result. When the new one wins, you will know why, in your own numbers, before anyone has priced it for you.

On "here."

The Pitch: "The Future of Radiotherapy is Here."

The Reality: The engineering is real, and seven serious institutions put their names on it. The timing is on Siemens Healthineers' own product page: "510(k) pending. Not available for sale in any market." That is where every new platform starts, and it says nothing against this one. Availability is also only the first date. After it comes the path into your building: the order, any vault work, installation, then commissioning and QA before the first patient, and each step takes time. So the date that matters to a program is a realistic first patient, not the launch. Put Accela on the roadmap as one line with a reassessment trigger: the change in case mix, or the replacement window, that would move it from watching to asking for a quote. Until one of those trips, the platform you treat on today is still doing the job it was bought to do.

With Heather Turner, RT(T), PMP

The $3 Million Floor Space: Why capital equipment upgrades fail before the riggers even arrive.

When an oncology program buys a new linear accelerator, the focus is almost always on the press release, the vendor negotiations, and the cutting-edge clinical capabilities. It is a multi-million-dollar investment designed to modernize your program and attract new patient volume.

Then the machine is installed, the ribbon is cut, and a painful reality sets in: the transition period dragged on weeks longer than planned, the department lost hundreds of thousands in displaced patient volume, and the clinical team is overwhelmed trying to adapt to new workflows while catching up on backlogged cases.

Having managed operations through major capital replacements, I can tell you that the success of an equipment upgrade isn't determined by the technology you buy. It is determined by how well you manage the operational void created during the installation.

Before you sign the purchase order for your next major technology upgrade, work to build your operational ramp-up plan.

Account for the "Displacement Cost": Removing an old machine and commissioning a new one takes months. If you are a two-LINAC department taking one machine offline, your remaining machine capacity drops by 50%. Do not assume you can simply squeeze those patients into existing hours. Plan early to extend operating hours, stagger shifts, or utilize weekend slots on your active machine long before construction begins.

Map the Patient Transfer Journey: When patients are shifted from their primary machine to another vault during construction, setup parameters, immobilization, and beam delivery can vary. Train your RTTs ahead of time on how patient transfers will be handled between vaults to eliminate day-of-treatment confusion, setup delays, and unnecessary stress for your clinical staff as well as your patients.

Phase the Clinical Go-Live: Don't attempt to launch a brand-new machine with a full, 40-patient schedule on day one. Build a phased ramp-up template: start with 20% capacity in week one, focusing on simple cases, and scale up incrementally as the team gets comfortable with the new console software and positioning hardware.

A new linear accelerator should be an operational milestone, not an operational crisis. When you plan for the physical and human realities of an equipment swap, you protect your bottom-line during construction and ensure a seamless go-live.

Before the next quote arrives, write the program's own map on one page.

Case mix, today and in five years. Disease sites, techniques and fractionation, including where stereotactic and hypofractionated work is heading in your referral base.

Volumes by site. Courses and fractions per machine and per location, and where the schedule is actually tight.

Service lines planned. What the program intends to add or expand in the next five years, with a date beside each one that has a date.

What each feature would change. List the headline features of any platform on the table, new or current, and beside each one write which line above it would move: capacity, technique, access or downtime. A feature that moves nothing on your page is impressive and not yet yours.

Then read the current machine and the new one against the same page, and sort what you find three ways: what you adopt now, what goes on the roadmap with a date or a trigger, and what waits for a later cycle. Once it is mapped, a launch is a line on your own roadmap, on your own dates.

Medsolve takes one document you already have and sends back one page finance can act on. No charge.