Changelog

What’s new in CBCTHub

Every release, told by what it lets you do. The big things first; everything else in the list below.

headline features: 35
improvements and fixes: 94

Sep 5 — Sep 5, 2026

The full Ricketts, and the profile that draws itself

OrthodonticsSep 5

The Ricketts analysis, complete: twenty-two rows and the VERT coefficient

Ricketts organised his analysis into six "problems" — dental, skeletal, dentoskeletal, aesthetic, determinant and internal structure — and the sheet now follows that order, which is part of the method because it says in what order a case is read. Eight measurements that did not exist were added: maxillary depth and height, cranial deflection, cranial length, porion location, ramus position, the mandibular arc and the mandibular corpus length, plus lower facial height measured the way Ricketts measures it, which is not the ANS-Me distance it gets confused with. The last of these needed three landmarks that are now placed — Xi, the centre of the ramus; Dc, the condyle neck; and Pm, on the symphysis — and two that are never placed by hand because they build themselves: the cranial centre and the facial centre. With the mandibular arc available, the VERT appears: Ricketts’ growth coefficient, a single figure printed beside the case headline that says whether the face grows forward — brachyfacial, deep bite — or downward — dolichofacial, open bite tendency. It is the average of five measurements and is printed only when all five are there: with four it does not give a number, it says which one is missing.

OrthodonticsSep 5

The tracing on its own, without the radiograph underneath

The cephalometric report now ends with a plate: the patient’s lines on white, with no radiograph underneath. It carries the bone outlines, the soft tissue profile, the reference planes, every landmark with its abbreviation, and a dozen values written where they are measured — the facial axis at the cranial centre, the mandibular plane where it crosses Frankfort, the interincisal angle between the two incisors. It is the sheet that gets shown and filed, and it does three things the radiograph with the tracing on top cannot: it photocopies without turning into a grey smudge, it overlays on the tracing from the first visit to see how much the patient grew, and it can be shown to someone without showing them a radiograph. The numbers are a selection rather than the whole table: twenty-two of them spread over a face collide with each other and cover exactly what you came to look at. And none is invented: a measurement that could not be computed is not written. With fewer than eight landmarks the plate is not produced, because a figure that does not look like a face, signed as "the patient’s tracing", is worse than not offering it.

ReportSep 5

The report grows to as many pages as it needs

The cephalometric report sheet was a single page, and whatever did not fit was not printed: with the full Ricketts analysis, the last measurements were drawn on top of the legal notice and the source list — where each norm comes from — was cut in half without warning. A sheet that asserts twenty-two numbers and cites eleven sources is worse than one citing none, because it looks complete. The report now takes as many pages as it needs: the table breaks between rows, each new page repeats the column names and a header with the patient and the analysis so a loose page can be identified, the legal notice appears on all of them, and each carries its number. Short analyses still come out on a single page. This applies both to the downloaded PDF and to the image that goes into the report.

OrthodonticsSep 5

The soft tissue profile draws itself, from the radiograph

The soft tissue profile now draws itself, from the radiograph. It used to be built by joining the ten soft tissue landmarks placed by hand — glabella, soft nasion, nose tip, columella, subnasale, lips, stomion, mentolabial sulcus and soft pogonion — so in practice it was almost never seen: automatic tracing places four of those ten, and since a missing point cuts the curve rather than bridging it, what was left was a five-millimetre line between subnasale and the upper lip. The app now finds the skin-against-air edge, the highest-contrast boundary on the whole plate, and draws the entire profile: forehead, nasal dorsum and tip, lips, chin and neck. It is drawn solid rather than dashed, because it is not an interpolation between points: it is the edge, pixel by pixel. It is computed in your own browser, in a tenth of a second, and the radiograph goes nowhere. It draws, it does not measure: the E line and the nasolabial angle still come from the landmarks you place, because a measurement has to be something you can point at. If the plate is cropped in front of the nose, or the profile cannot be made out, it draws nothing rather than inventing a curve, and the "Auto profile" button turns it off.

Also in this release

Sep 5The soft tissue profile line comes out smoother. It is extracted row by row from the radiograph, so every point was pinned to the pixel and the curve drew jagged, like a saw blade — and a jagged soft tissue profile does not look like a face. The zigzag drops by almost half: the points that get saved are no longer taken by skipping one in four but by averaging those in each stretch, which is what cancels the stepping. What is lost that way measures less than a millimetre and a half, that is, nothing anatomy has: the subnasal and mentolabial sulci stay intact, which was the risk of over-smoothing.
Sep 5The report no longer contradicts itself about the vertical pattern. The mandibular plane angle was computed two different ways — one for Tweed’s FMA and another for the Ricketts and Downs rows — and as long as the angle stayed under 90 degrees both gave the same result, so they coexisted unnoticed. Past 90 they diverged, and the same sheet then said "Vertical pattern: hyperdivergent (dolichofacial)" at the top and "hypodivergent, brachyfacial pattern" three centimetres below. It happens above all on unreviewed automatic tracings, where Menton can end up behind Gonion. All three rows now use the same computation, and an impossible tracing prints an impossible number — outside every norm, with three asterisks — instead of one that looks like a patient’s.
Sep 5Two rows that could read backwards have been fixed: Ricketts’ facial depth and Downs’ IMPA. Both were computed as the acute angle, which runs from 0 to 90 degrees, and both sit right beside 90: past that point the number folded over and started coming down again. In practice an increasingly prognathic chin printed as increasingly retruded, and a lower incisor at 103 degrees — heavily proclined — came out as 77 with the word "retroclined" next to it. Both now use the same computation as their twin measurement, which was correct all along: facial depth matches Downs’ facial angle, and Downs’ IMPA matches Tweed’s, which is what they should always have done. If you have filed Ricketts or Downs reports for Class III patients or for heavily inclined lower incisors, they are worth reopening.
Sep 5Every landmark placed on the plate can be adjusted again, whether the chosen analysis asks for it or not. Automatic tracing places nineteen landmarks at once and several do not belong to the analysis you are using — Steiner asks for neither Porion nor Orbitale, and the four soft-tissue ones only come in with the profile turned on — so they were drawn on the radiograph but left out of the list, and correcting one meant hitting it with your finger, with no name, no description of where it belongs and no loupe. They now appear at the end of the queue and in the grid, under "Other placed landmarks", and the arrows reach them like any other.
Sep 5The loupe is now there when giving the plate its scale, too. It only existed in the landmark step, so calibrating — two clicks on the cephalostat ruler marks — was done with the naked eye, which was exactly backwards: that distance yields the millimetre every one of the report’s fifty-six measurements is then computed with, so a pixel of error there spreads across all of them. The 4× panel now follows those two clicks as well, and when the cursor is off the image it focuses the ruler point you just placed.
Sep 5A lateral ceph can now be traced automatically and corrected by hand afterwards. A model proposes where each landmark goes and the points are placed straight away: the lines are drawn, the numbers appear, and the orthodontist drags whichever ones are off. The ones nobody has touched yet look different — hollow and amber — and the panel says how many are left; dragging one takes it off that list on its own. While any remain unreviewed the printed sheet says so, with the count and the model they came from, so a filed report always states where its landmarks came from. The mark disappears entirely once they have all been reviewed, because by then the tracing is as hand-made as any other. It is help with the repetitive work, not a diagnosis: the model gets things wrong, which is why every landmark stays editable.

Sep 3 — Sep 4, 2026

The lateral ceph, end to end

AnalysisSep 3

Cephalometrics is traced on the plate, and stays with the study

You place the landmarks on the very lateral ceph you are looking at, with the analysis panel alongside: pick Steiner, Ricketts, Downs or any of the nine, the panel asks for the next point in turn, and the measurement table updates as you go. A loupe follows the cursor so Nasion can be placed to the millimetre. When you are done, "Save to Analysis" keeps the tracing with the study: the radiograph with the lines drawn on it, and the results sheet with the table and the chart. The first one reopens with your points exactly where you left them, so you can pick a case up weeks later or compare today’s tracing with the one from the first visit. And it no longer opens in a separate window: you work inside the viewer, on the plate, without going anywhere.

PanoramicSep 3

Measure, point and annotate on any radiograph

The radiograph viewer now has measure, arrow and note, with the same gestures as the CBCT viewer: two clicks for a line or an arrow, one for a note; drag them by an end or by the body, delete with the X that appears on hover, change an arrow’s colour with a double click. Everything is saved with the study, image by image: what you marked on the first plate does not show up on the second. Nothing is invented about scale: a radiograph that arrived as a file lost how much each pixel measures, so when you pick the ruler the viewer asks whether you want to calibrate. Measure something of known size — on a lateral ceph, the cephalostat’s own ruler, which is magnified exactly as the patient is — type what it really measures, and every measurement on that image switches to millimetres at once. If you would rather not calibrate, it measures in pixels and the label says so.

ExamsSep 3

The radiograph viewer has Analysis and Report, like the CBCT one

A radiograph now opens with the same header and the same tools as a CBCT: the rail on the left with the pointer, brightness and contrast, zoom and pan; up top, the study, the patient aligned to the right, and the Analysis, Report and Cephalometrics buttons. The Analysis is the same card bank: you save the image you are looking at — with the brightness and contrast you set — reorder by dragging, name each card and export the lot as a PDF. And brightness and contrast are adjusted by dragging, as in any DICOM viewer, instead of three sliders taking up a permanent row to say 100 %.

OrthodonticsSep 4

The landmarks join up: the tracing now draws the profile, the mandible and the maxilla

The landmarks you place are no longer loose dots: they join into the outline of the soft tissue profile, the mandible, the maxilla and the cranial base, each structure in its own colour. It earns its place three times over. It shows a landmark that sits wrong, which is the hard thing to catch: twenty-nine dots on a radiograph have no shape, and a Gonion three centimetres too high looks exactly as good as one in the right place — joined to the rest it throws a bulge into the lower border of the mandible that you cannot miss. It explains the case to the patient: forty numbers in a table do not show why the chin is set back; the drawn profile does. And the drawing goes into the plate saved to Analysis and into the sheet you download. So that the profile has the shape of a profile, three further landmarks can be placed — soft tissue nasion, stomion and the mentolabial sulcus — in a list of their own and entirely optional: no analysis asks for them and they do not count towards the tally. The outline is dashed on purpose, because it is the curve through the landmarks and not the edge of the bone; and where a landmark is missing the line stops there rather than skipping it, which would draw a mandible with no angle. Toggle it with "Outline", next to "Lines".

OrthodonticsSep 4

The tracing now measures overjet and overbite, and every norm has been checked

Two measurements every case is judged on were missing from the summary and are now in it: overjet and overbite, with their 1–3 mm norm. Wits asks for two new landmarks — the upper and lower premolar cusps — because it is taken on the functional occlusal plane, not the bisected one Downs uses; the tracing carries both planes and each analysis takes the one that belongs to it. Incisor and molar abbreviations read U1, L1, U6 and L6: the bars used before sat on the wrong side, so "upper" appeared where the landmark was the lower one. The McNamara and Ricketts norms, along with the gonial angle, were checked against the published tables and corrected where they did not match. And a deviation is no longer flagged all the same way: the sheet tells mild from marked by how far the value sits from the norm, and when the patient’s sex is missing it says so, instead of colouring green a comparison that never happened.

AnalysisSep 4

The results sheet downloads as a real PDF, not a picture

Until now the file was a screenshot wrapped in a PDF: fine on a monitor, poor on paper, with type that blurred the moment you zoomed and nothing you could select or search. The PDF now carries text as text and lines as lines, so it prints crisply at A4 or any other size, a figure can be copied out of the table, and a measurement can be found with Ctrl+F. The radiograph is still an image, because that is what it is, but the tracing over it, the measurement table and the bar chart are properly drawn. It is also a good deal smaller, which helps when the sheet goes out attached to an email.

Also in this release

Sep 4The real-distance field, when giving a radiograph its scale, no longer shows an example number. It said "10" in grey, and a number inside a field reads as a value already typed: you had to look twice to notice the field was empty. It now says "How many millimetres is it?", which cannot be mistaken for entered data. This is the step every millimetre in the report depends on.
Sep 4Each analysis is now compared against its own author’s table, and Downs, Ricketts and McNamara are complete. An analysis is not a list of angles: it is one author’s table measured on one author’s sample, and the authors do not agree. Downs was using Tweed’s mandibular plane — 25 ± 5 instead of 21.9 ± 3.24 — and with that nearly twice-as-wide deviation a patient at 28 degrees, plainly dolichofacial for Downs, came out within norm; and Steiner’s interincisal angle, four and a half degrees below Downs’, on the measurement that decides whether incisors can be retracted. Ricketts was using Downs’ facial angle, which is fixed, when Ricketts’ facial depth grows 0.3 degrees a year: in an adult that is more than two deviations. Downs goes from seven measurements to its ten — adding incisor to occlusal plane, incisor to mandibular plane and upper incisor to A-Pogonion — Ricketts gains facial taper, palatal plane to Frankfort and upper molar to the pterygoid vertical, and McNamara gains upper incisor to A-vertical, its row about the teeth, which was missing. Rows measuring the same angle against different tables say whose table each one is.
Sep 4The upper molar to the pterygoid vertical is now compared against the patient’s age. Ricketts’ norm for this measurement — the one that says whether there is room to distalise or extraction is needed — is not a number but a rule: the distance should be the patient’s age plus three millimetres. With a date of birth on file it applies itself, and it stops at 21 millimetres because growth stops: with no ceiling, at forty the norm would ask for 43.
Sep 4The loupe moves off the radiograph and into the panel, below the instruction. Fixed at the bottom right it covered two hundred pixels of film, and the bottom right is exactly where the chin and the soft-tissue profile fall: Pogonion, Gnathion, Menton and the lips sat under the box precisely when it was their turn to be placed. In the panel it covers nothing and sits next to what you are reading, so the eye goes from the text to the detail without crossing the screen.
Sep 4The cephalometric sheet now opens with the case summary. Above the forty-row table sit the headline — "Class III · Brachyfacial" — and three columns with what gets read first: skeletal class and where each jaw sits, the vertical pattern, overjet, overbite and incisor inclination, and the soft-tissue profile; each line with its grade — mild, moderate, severe — and with the abbreviations of the measurements it came from, so you can drop into the table and check it. Nothing new is computed behind it: it is the same table said in words, on the same scale as the asterisks, so the summary and the rows cannot disagree. And when a landmark is missing the finding simply does not appear, rather than being half-asserted.
Sep 4Placing cephalometric landmarks is now one at a time, with a drawing of where each one goes. The panel used to show thirty tiny pills with abbreviations, and above them, in small type, the only thing that mattered: which one is next and where it is. Now the current landmark takes the space — the name large, the anatomical description, and a small normal tracing with the landmark lit in place, which is what you need to find Articulare or Porion the first time — with arrows to move back and forth through the queue and a progress bar. The full grid is still there, folded away, because going back to one particular landmark to correct it is what happens at the end. The analysis is chosen at the top, large, and says what it is for.
Sep 4The loupe no longer chases the cursor: it sits fixed, and larger, at the bottom right. Following the hand it was close by, but moving — while nudging a landmark by a pixel, the thing you are looking at slides with your hand and your eye chases the box instead of studying the detail. In a fixed place you learn once where it is. And it no longer goes dark between two clicks: with no cursor over the image it focuses the landmark that is next, so arriving from the panel it is already showing where to look. It fades out on its own if the cursor lands on it.
Sep 4Each row of the sheet now has numbers on its scale. It was a mute ruler: you could see the value fell to the left, but not to the left of what, and to know whether that was two degrees or fifteen you had to go back to the norm column and do the arithmetic. The mean and two standard deviations either side are now printed on it — "91.8 on a scale running from 84 to 92" — and the row reads on its own.
Sep 4The Jarabak polygon is complete. The analysis showed the S-Go / N-Me ratio and the two facial heights, but was missing two of its four angles — the saddle angle (N-S-Ar) and the articular angle (S-Ar-Go) — both computed from landmarks already being placed. With them comes the Björk sum, the first thing anyone using this analysis looks at: 396° ± 6 is a balanced face, below it growth is horizontal, above it vertical.
Sep 4A 5 mm overjet no longer prints as a marked deviation. Overjet, overbite and the Jarabak ratio were written as ranges — "1 to 3 mm", "62 to 65 %" — and the sheet used half the band as if it were a standard deviation: a 5 mm overjet, an everyday Class II, came out with three asterisks and the word "marked". They are now compared against a real standard deviation and that same case shows one. And where no standard deviation has been published, the sheet prints no asterisks at all: the unit of a banded norm is not a standard deviation, and saying it with the same symbol claimed something unknown. Every norm states where it comes from, and the two with no paper behind them say so.
Sep 4Ricketts convexity is corrected for the patient’s age. Ricketts defined it at nine years old, with a correction of 0.2 mm less per year thereafter. It was being compared against the nine-year-old figure of 2 mm at any age, so a forty-year-old with a straight profile came out three deviations off: "markedly concave profile" in a healthy person. With a date of birth on file the norm shifts on its own; without one it stays at the reference age and the measurement says so.
Sep 4The results sheet can be read in print. It was composed at 5.8 points of body text, 4.8 for each measurement name and 4.3 for the legal notice: for a document that gets filed and shown to the patient, that is below the reading threshold. Body text is now 8.9 points and the footer 6.7, with the columns rebalanced so text does not spill into the next one, and the closing notice is no longer cut mid-sentence. The sheet also carries a new line with the study and the image it came from — received on its own there was no way to tell — and saving it to Analysis no longer recompresses it: it was text on white turned into a JPEG, leaving the norm sources illegible in exactly the card that travels through the shared link.
Sep 4Cephalometrics can be opened on a narrow screen. The options panel was a fixed 396-pixel column beside the image, so on a 390-pixel phone the image was left zero pixels wide: clicks landed nowhere and nothing could be marked. Below a thousand pixels the image now goes on top and the panel underneath. And if the screen really is small it says so: at that size one pixel is close to one millimetre of the patient, so a finished tracing can be reviewed, but placing landmarks accurately needs more width.
Sep 4What you mark no longer disappears silently when a save fails. The viewer waited 1.2 seconds and sent the changes; if that request failed — no network, or a flaky connection — the work was discarded and the panel carried on showing "21/21" in green. What is pending is now held until the server confirms it, and when it does not, a bar appears at the top with a retry button. And if your account can view the study but not change it, you are told on the first attempt instead of marking for half an hour for nothing.
Sep 4Correcting a landmark no longer moves it three millimetres at once. When you started dragging, the point jumped to wherever the cursor was: since you can grab it from slightly off to the side, grabbing Nasion and moving a hair shifted it by up to three millimetres — and at screen scale that jump is invisible. The offset you grabbed it with is now kept, so dragging moves it instead of teleporting it. On top of that, "Clear tracing" asks for confirmation before taking twenty-nine landmarks with it, and on a tablet a second finger or an incoming call no longer leaves the point stuck to your finger.
Sep 4Calibrating a radiograph’s scale now warns you when the result cannot be right. All it asked for was two distinct points and a positive number, so two clicks three pixels apart declared as ten millimetres produced a scale thirty times too large, and every millimetre measurement was printed that way in the PDF without a single flag. A minimum baseline is now required — below it, one pixel of error changes every measurement in the study — the computed scale is shown before you apply it, and if it cannot belong to a radiograph it says so. And calibration can now be removed, not only redone: "recalibrate" used to clear the ruler but leave the previous scale active.
Sep 4Five cephalometric measurements could print a wrong number that looked normal, and no longer do. Overbite was measured against a vertical taken from the two molar cusps, which occlude and are almost never one above the other: half a millimetre of offset when placing them dropped it from 2.5 to 1.1 mm, and with both in the same spot it printed 0.0 — which reads as an open bite. It is now measured perpendicular to the occlusal plane. The occlusal plane to Frankfort and the three incisor angles to N-A, N-B and A-Pog did not tell one side from the other: an incisor retroclined by 30 degrees printed the same number as a normal one, and the more it was retroclined the higher the number went. They are signed now, so a negative value means what it says. And when a line is left with no direction — premolars placed on top of the molars, point A on top of Nasion — the row is left blank instead of printing 0.00 mm, which in Wits reads as a perfect skeletal relationship.
Sep 4A growing patient is no longer silently compared against adult norms. The date of birth was asked for, printed on the sheet, and used in no calculation at all: a healthy nine-year-old got five pathological rows and a Class II that does not exist out of the McNamara analysis, purely because the mandible is thirty millimetres shorter than an adult’s, exactly as it should be at that age. The eight size measurements are now still measured but not compared when the patient is under sixteen: the norm column reads "adult norm" and the sheet says so. With no date of birth everything works as before, because with no age there is nothing to assert.
Sep 4The sheet’s PDF no longer prints question marks where the male and female symbols belong. When sex is not chosen, the norm column shows both — the male one and the female one — and told them apart with ♂ and ♀, which do not exist in the PDF fonts: it came out as "Pog ?N ? 0.2 mm ? -1.8 mm", two contradictory norms with no way of telling which was whose. They are spelled out now. And three McNamara abbreviations that used the perpendicular symbol and an en dash are written with characters the PDF actually has.
Sep 4The patient’s email can now be typed when you upload the study and corrected later, both under "Edit" and in the send form itself: it is the same address in all three places. It is optional. Until now it only filled itself in the first time something was sent, so if the patient spelled it wrong that is where it stayed: the send form showed it in a box you could not touch and, even if you typed another one, it still sent to the old one. A malformed address is flagged as you type — half an address is worse than none: the send is taken as done and the link goes nowhere.
Sep 4When you upload a bitewing, the four slots are now named by sector: right molar, right premolar, left premolar and left molar, in the usual mounting order — the patient’s right on top. They used to read "upper right", "lower left" and so on, which describes a periapical series: a bite-wing film carries the upper AND the lower crowns on the same image, which is the whole point of it. With the old labels the four films went into the wrong slots, and whoever read them later was looking at the wrong side.
Sep 4A bitewing or a full periapical series now opens with every image on screen instead of on the first one. They are not several radiographs: they are one examination spread over four or fifteen films, and the diagnosis comes from comparing them. Until now the first thing to do was always to hit "view all", and anyone who did not know that button existed sat looking at one film of a four-film examination. A single periapical still opens one at a time, because there each film is a different tooth. And the green chip up top now reads "Periapical series" when that is what it is; it used to say plain "Radiograph".
Sep 3The patient’s name now appears in the radiograph viewer’s header, hard right and on its own, with no study name under it: which radiograph you are looking at is already spelled out by the green chip on the left. It also travels through the shared link: whoever received a radiograph used to see the study name and no patient at all.
Sep 3The "new support reply" notification now opens that conversation directly, not the ticket list. With one ticket it made no difference; with five you had to guess which one held the reply. It works for notifications you already had stored, too.
Sep 3A study’s details panel now has the delete button at the bottom of Actions. Before, you had to close it and find the three-dot menu back in the list. It sits at the end, set apart from the rest — everything above is done daily, this is done once and cannot be undone — and it asks for confirmation spelling out what goes: the DICOM files, the report, the extra scans and the links already shared. Owner and administrator see it, same as in the list.
Sep 3Support tickets now read as a conversation. Your tickets are in the Help Center, and each one opens to show what you wrote and what we answered, with screenshots from both sides. The bell used to say "new support reply" and lead to a page with no reply on it: the answer only existed in the email, so anyone who deleted it or had it in spam was left with nothing.
Sep 3And support can now send you screenshots in a reply. They show inside the email, not just as an attached file, and they also stay in the conversation in your dashboard. Explaining "the three-dot menu, the option below" with an arrow drawn on it takes a second; in words it takes three paragraphs and a follow-up question.
Sep 3With a single study in the list, the three-dot menu now shows Delete. The list is a rounded box that clips whatever overflows it, and with one row the menu did not fit: you saw Download and that was it. It was reported as the free plan not being able to delete studies, and it was not the plan — it was the clipping, and it happened to anyone with one study, which is exactly what a new user has.
Sep 3In TMJ, dragging with the zoom tool selected no longer changes the brightness. The viewer could be left believing Shift was held down when the keyboard missed the release — switching windows, or the layout rebuilding mid-gesture — and since brightness lives on Shift + drag, it took over the drag whichever tool was selected. Now, on mouse down, the viewer checks whether a key is actually held.
Sep 3The framing of the TMJ slices is saved even when the other condyle is selected in the header. With both joints on screen, the viewer only saved it if the side chosen in the header was the one with a traced axis: trace one while the other was selected and your adjustment was lost on return.
Sep 3In TMJ, opening both joints no longer leaves the axial zoomed in and cut off at the sides. The axial is a tall narrow column when both joints are shown and a wide short one when only one is: the framing you set in one view is no longer imposed on the other, and each remembers its own. It matters for tracing the condylar axis, which has to be seen whole, from the lateral pole to the medial one.

Sep 2 — Sep 2, 2026

What is around the implant

ImplantsSep 2

The sinus, the nasal floor, the adjacent root and the implant next door

The viewer could measure one thing: the inferior alveolar canal. And since the canal is mandibular, an implant in the maxilla got nothing measured at all — the row read "—" and that was that, even with the sinus floor a millimetre from the apex. You can now trace the sinus floor, the nasal floor and an adjacent tooth root with the same gesture used for the canal, and say which is which in the annotations panel. Each has its own colour, its own threshold — 1 mm under the sinus and the nasal floor, 1.5 mm to the adjacent root, the usual 2 to the canal — and its clearance shows up in the list, the details, the report and the plan PDF. Entering the sinus is worded for what it is, "enters the sinus", not as a mistake: after a lift that is exactly the intent. And the spacing between neighbouring implants is measured too, the other figure that decides a plan: the rule quoted is 3 mm, and it is measured along the whole body rather than only between platforms, because two implants can leave three millimetres at the top and touch at the bottom.

ImplantsSep 2

Tooth numbers, in the notation used where you are

The same tooth has a different name depending on where you work: 11 in FDI, 1.1 as it is written in Chile, and 8 in the Universal notation used in the United States. You now choose which one you see, right in the chart where you assign the tooth, and it applies across the whole viewer at once: the list, the details, the label on the slice and the plan PDF. The chart itself does not rearrange — a tooth is in the same place in the mouth either way — only the number written in each box changes. And the FDI number is still what gets stored: it is what decides the quadrant and finds the crown shape, so a case opened in one notation and in another is still the same case.

DashboardSep 2

The data protection agreement, in your account and on the record

Settings has a new section, Data protection, with the full data processing agreement: what CBCTHub may and may not do with your patients’ data, who else is involved and for what, how a breach is notified and what happens when the service ends. You read it right there — no PDF to download in order to find out what it says. Accepting it records the date, the version and the document fingerprint, and that record can be printed or saved. It matters because your practice is the controller under the law and CBCTHub only the processor: this agreement is exactly what the regulation requires you to hold with any provider. And if the provider list ever changes, the document moves to a new version and you are told here — nothing you already accepted is edited quietly.

Also in this release

Sep 2The home page loads five times lighter: the viewer video went from 5 MB to 1 MB and no longer downloads until it is about to be seen. It shows most on a phone using mobile data.
Sep 2The home page shows the viewer before saying anything else: the product video sits right under the buttons, and the figures under the video. It used to take a screen and a half to see anything.
Sep 2The home page opens with one short sentence saying what CBCTHub is, and under it four figures: active users, imaging centers using it, compatible scanner brands, and how current it is.
Sep 2The ten names in the compatible-scanner strip are links now: each one goes to its page, the one that explains how to export from that machine. They were plain text.
Sep 2The home page says in one sentence what CBCTHub is, in the language of whoever opens it; it names Romexis, Invivo, OnDemand3D and DEXIS outright — each linking to its own comparison — and it shows who builds it: name, specialty and a link to the profile. Plus four figures anyone can check on the site itself, the date of the latest release among them.
Sep 2Each implant’s slice in the plan PDF now comes with the crown and the abutment drawn on it, just as on screen. Drawing only: it adds no measurement.
Sep 2The abutment and implant-details texts say the same thing in one line: whoever reads them already knows implants.
Sep 2The home page answers three more questions: what CBCTHub is, what you can do with a scan, and how it differs from each brand’s desktop program. They were written for search engines only — invisible, and in Spanish even on the English and Portuguese sites; now they are readable, each in its own language.

Sep 1 — Sep 1, 2026

Implants, in one list

ImplantsSep 1

The implant list and its report

When there are implants, an Implants button appears on the panoramic, and with it the table of them all: tooth, model, size, distance to the canal and density, one under the other. From there you jump to any of them — even one in another part of the arch — change its tooth or its size, and switch it off on screen to see the bone underneath without deleting it. And it exports to PDF: the panoramic with the table on the first sheet, and each implant’s detail with its slice after it.

ImplantsSep 1

The simulated crown: anatomical, joined to the implant, and now on the slices too

The crown that appears over an implant once you assign it a tooth now uses the piece from the library: a molar’s cusps, a premolar’s profile, each tooth’s real shape. The library had been there all along, but its files never actually loaded and the viewer always showed the fallback — the formula-generated shape. Because that fallback works well, nobody could tell the library was never being used. And the crown no longer lives only in 3D: it is drawn on the cross-section too, which is where you decide whether there is buccal bone left and whether the implant axis comes out where it should. It is drawn where the slice CUTS THROUGH it, not as its shadow: if the crown is three millimetres away, that slice shows nothing. And it now rises from the implant platform instead of floating above it: library teeth are laboratory teeth and their neck is a natural tooth’s — a lower first molar opens eleven millimetres over a four-and-a-half millimetre platform — so the emergence profile between the two is generated using the diameter of the implant actually placed. The neck is no longer a straight cut either: it rises toward mesial and distal, like a tooth’s cervical line.

ImplantsSep 1

The abutment, between implant and crown

A whole body was missing: the crown rose straight out of the implant platform. The abutment is now drawn — the gingival portion, the shoulder and the tapered post that runs up inside the crown — in 3D and on the cross-section too, which is where you see whether the margin sits above or below the bone crest. The abutment’s height decides where the crown seats, and its angulation is what lets a tilted implant carry an upright crown: both are judged against the bone. From the implant list you open the abutment panel and set four of them: gingival height, margin diameter, post height and angulation. The other five follow from these and from the implant diameter. The abutment appears as soon as the implant is placed — no tooth needs to be assigned, because an abutment screws into the implant, not into a tooth. The crown does need one, and it seats on the post: at 0° — a straight abutment, a Ti-base — the crown comes out with whatever tilt the implant has, and you raise the angle until it sits where it should. That number is the angled abutment to order. Double-clicking the implant opens its details — implant, abutment and crown on one screen — and there you pick the abutment family: Ti-base, straight, angled 15° or 25°, and multi-unit straight, 17° or 30° for full-arch. Each one fills in its measurements and you fine-tune from there. These are each family’s typical measurements, not any manufacturer’s catalogue.

ExamsSep 1

Downloads are visible, can be stopped, and say when something is missing

While an exam downloads, a card at the bottom shows how much has come down, how much is left, and it stays on screen even if you close the exam card or open the viewer. You can cancel from there. If you try to leave the page mid-download, the browser warns you before you lose it. And if a file did not make it, it now says so: before, the exam came out anyway, with the right names and a file or two missing, which is the worst way to fail because it looks complete. The warning also travels inside the ZIP, for whoever opens it the following week.

ExamsSep 1

A scan downloads as a scan, not inside a ZIP

An intraoral scan or a radiograph now downloads as the file it is, with nothing around it: whoever receives it opens it straight in their planner. A CBCT still comes as a package, because it is not one file but a package: it carries the DICOMDIR and the folder tree from the CD, and that is how the person opening it expects to find it. The difference shows most on a phone, which is where the link a patient receives usually gets opened.

Also in this release

Sep 1The implant plan PDF fits on a single sheet when there are only a few, the panoramic is bigger, the implant model is no longer clipped, and the density no longer repeats the word heading its own column.
Sep 1The downloaded file’s name keeps accents and ç: “João Gonçalves” no longer arrives as “Joo Gonalves”.
Sep 1The toolbar is chosen separately for each place: your viewer opens with the side column, and the shared link and the embedded viewer with the bar on top. All three are set in Settings.
Sep 1Downloading no longer asks for a format. The question meant inspecting the exam over the network before starting, and on a slow connection the button simply did nothing: it will come back once that wait has a limit.
Sep 1When an exam arrives compressed and cannot be converted, the exact format is named along with how to ask for it uncompressed.
Sep 1On the panoramic, zoom now works by dragging and anchors on the point you are looking at, with a hand tool to pan — the same as in Explore.
Sep 1Slice thickness uses the same control in the same place across Explore, Panoramic and Oblique, and the bottom bar is identical in all three.
Sep 1“Reset” on the panoramic brings back the standard curve and clears the view.
Sep 1This very page: the public site now has a section with what keeps changing, in all three languages.

Aug 27 — Aug 31, 2026

Cephalometrics, saved views and recorded readings

OrthodonticsAug 27

Full cephalometric analysis inside the viewer

Mark the landmarks on the lateral cephalogram and get the nine standard analyses — Steiner, Ricketts, McNamara, Tweed, Downs, Jarabak and others — with every value compared against its norm and explained in plain language. It ends in an A4 sheet ready to print or attach, without leaving for another program.

CollaborationAug 28

Two people in the same exam

What one window saves shows up in the other, so no single machine holds editing rights any more. The centre also decides whether whoever receives the shared link can save what they mark; if not, they are told their marks stay on their own screen.

PanoramicAug 28

Walk the arch with the slices

Cross sections are spaced by real millimetres and form a block you drag along the arch by hand, with the wheel or with the slider. Stretch it from the edges to cover more area without losing detail.

Explore · ObliqueAug 29

Serial slice grid

Four, nine or sixteen contiguous slices at whatever millimetre spacing you choose, each with its ruler, its slider and the reference lines of the other planes. The whole grid drags into the report or the plate as a series, one frame per slice.

AnalysisAug 30

Saved views: get back to where you were

Save the viewer’s full position under a name — “Tooth 1.2” — and return to it with one click, without losing the work in progress. The Analysis panel becomes a grid you order by hand, with a link so a colleague opens exactly that view.

ReportAug 30

Recorded readings, voice and path

Record the spoken explanation together with everything done on screen. Whoever receives it plays it, pauses it and carries on from wherever they like. It comes transcribed and with per-tooth markers to jump to a finding, exports to video, and is sent by email with a link that expires.

Also in this release

Aug 28A button flips the sagittal view left to right, with the ANT and POST labels following the flip. It is saved with the exam.
Aug 28The shared link and the embedded viewer honour the centre’s preferences, including where the toolbar sits.
Aug 29Implants, measurements and notes on the panoramic stay pinned to the anatomy and remain visible even when the slices are far away.
Aug 29The PDF report gets its accents back, and its footer shows the clinic’s details instead of the software’s name.
Aug 29Drag the CBCT folder onto the exam list to upload it. Drop JPG or PNG images instead and they come in as radiographs, asking which kind they are.
Aug 29The exam search moves into the header, and the storage warning arrives from 80 % with how much is left and how fast it is filling.
Aug 30The comparator gains “Align here”: put both studies on the same anatomy once and from then on they move together.
Aug 30Choose what goes with the exam when sharing it — report, captures, saved views and readings — with the count of each.
Aug 31The 3D implant now looks like a real implant, and once a tooth is assigned a simulated crown appears showing where the screw would come out.

Aug 16 — Aug 26, 2026

Orthodontics, TMJ and automatic upload

ReportAug 16

Plate composition

Build plates by picking the grid in a visual selector, with quick fill or free composition, undo, drag to reorder, alignment magnets and a reading zoom over the sheet. Captures include the measurements, arrows and notes made on the slice.

ReportAug 16

Clinical report inside the viewer

Dictate it by voice and the text appears live, with the option to arrange it into clinical prose. Drag slices from the study into the body of the text and adjust each image’s size, brightness, contrast and crop. The PDF comes out with the layout, the images and the study details.

TMJAug 18

Both TMJs on one screen

With a single scan covering both joints they appear side by side, the reference axial in the middle and its slices below, instead of switching between right and left. The radiologist declares which side each acquisition is, and that travels to the report and the plates.

CBCTHub LinkAug 21

Machines upload the exams on their own

The centre’s computer is linked with a six-character code — no passwords, no configuration — and from then on it uploads studies automatically. Settings shows which machines are connected, when each last uploaded, and lets you cut their access.

ImplantsAug 23

Planning with real measurements

Place the implant on the panoramic and the viewer reports the distance to the mandibular canal measured from the implant’s edge, the corresponding tooth, and the density of the area. Implant and canal also appear in the 3D view, to check how much bone is left and whether it breaks through the cortical plate.

OrthodonticsAug 24

Airway and intraoral scan

Measure the airway by volume and minimum cross-sectional area, with the height profile showing where it narrows, then export the image or take it to the report. The patient’s intraoral scan can also be uploaded and comes out aligned over the CBCT, next to the slices.

Also in this release

Aug 17Zoom and framing stay in sync across the slices of one series.
Aug 17The report opens floating or expanded, with an adjustable divider.
Aug 17A signed report can no longer be edited: to correct or extend it, an addendum is added with its own date and signer.
Aug 18A tooth selector, and sending the report to the referring dentist as a step in the flow.
Aug 18A density tool measures the grey value of a region in Explore and Oblique, with the caveat that in CBCT these are not true Hounsfield units.
Aug 18Lasso cropping in the 3D view: circle what you want to keep, and the crop carries over to the slices too.
Aug 18What the patient downloads looks like the centre’s original CD again, with the disc index and its companion files, so it imports into other programs.
Aug 18Measurements can be edited without switching the tool on, and the number no longer covers the line.
Aug 18The panoramic grid is chosen graphically and the preference is remembered.
Aug 18The volume can be cropped from Explore, and the crop is shared with the panoramic.
Aug 19The mandibular canal is traced as one continuous path across every slice, drawn onto the panoramic and corrected by dragging its points.
Aug 19An optional toolbar in a side column, in Explore, Panoramic and Oblique.
Aug 19A slice thickness control on the panoramic cross sections.
Aug 21The report says when it could not save, instead of claiming it did.
Aug 21Arrow and Note can be drawn and edited on the cross sections too, and moved by dragging.
Aug 23Brightness and contrast set by hand are saved with the exam.
Aug 23Automatic brightness and contrast no longer blows out the teeth or flattens the soft tissue.
Aug 23In the 3D view bone can be shown translucent, and the marking brush stays anchored to the bone as the model turns.
Aug 24The fourth quadrant in Orthodontics chooses what it shows: 3D, panoramic or lateral cephalogram.
Aug 25On changing plan, the customer gets an email in their own language with what the new plan includes.
Aug 25The 3D view opens facing front, and the model can always be rotated.
Aug 26Fixed the bug that made the viewer open blank when entering an exam.

Aug 1 — Aug 15, 2026

The report and the plates

SharingAug 11

Embedded viewer with the centre’s brand

The viewer a centre publishes on its own site shows its logo and its domain in a single top bar, alongside the exam’s tabs. Patient and dentist see the centre’s brand, not ours.

ExamsAug 14

Each acquisition keeps its own work

In exams with more than one acquisition, the panoramic curve, the annotations and the framing are stored separately for each and no longer overwrite one another. In TMJ, the both-sides view loads the volume belonging to each side.

ReportAug 15

Plate assembly, rebuilt

Drag images onto the sheet from any view without giving up measuring, annotating or scrolling through slices. Frames move, grow, are added and removed within each exam, without touching the template of every other report. The preview shows the sheet exactly as the PDF will come out.

DashboardAug 15

Filters in the exam list

Next to the search box, filter by date — today, 7 days, 30 days, this year — by study type and by referring professional. Only the filters that make sense for the exams on file are shown.

SharingAug 15

Share from the viewer

Without going back to the exam list, the study’s owner generates the link, sends it by email or copies the embed code, all from the viewer’s own menu. It honours anonymised mode.

Also in this release

Aug 5Exam downloads keep the original CD’s filenames or add the .dcm extension, so they open without renaming anything on Windows and in viewers like Horos, RadiAnt or OsiriX.
Aug 5Support tickets always send again, and confirm it with a clear message.
Aug 13Expanding a view actually enlarges the image and keeps the zoom, the pan and the orientation. The reference lines no longer vanish.
Aug 13Arrows, circles, notes and angles drawn on the 3D view stay anchored to the model and follow it as it rotates and zooms.
Aug 13The mandibular nerve tracing and the notes also show on the cross-section thumbnails.
Aug 14The magnifier turns itself off when another tool is picked, and the mouse wheel changes slice again in the grid and on the panoramic.
Aug 15Removed the flicker of the reference lines while scrolling slices, and the green flash when collapsing a view.
Aug 15Analysis and Report can no longer be open at once, which used to throw the screen out of alignment.
Aug 15The PDF preview sits above the whole viewer, with a visible close button next to the document.

Something missing?

Much of this list came from what the centres using CBCTHub every day asked us for.

Tell us what you need