Latest / Deep Dive Ortho — Orthopaedic Surgery, AI & Digital Health Podcast. / Ellipse Glenoid Version: Friedman Method Refinement

Transcript

Ellipse Glenoid Version: Friedman Method Refinement · Deep Dive Ortho — Orthopaedic Surgery, AI & Digital Health Podcast.

• 0:00
  Welcome to the deep dive, where we really plunge
• 0:03
  into the sources, the articles, the research,
• 0:05
  and pull out those key nuggets of knowledge just
• 0:07
  for you. Today we're diving into something that
• 0:10
  sounds maybe a bit niche, but it's actually incredibly
• 0:13
  vital in modern medicine. We're talking about
• 0:15
  precision, absolute precision in shoulder surgeries.
• 0:19
  Imagine for a second you're maybe facing total
• 0:23
  shoulder arthroplasty, TSA, they call it. It's
• 0:26
  this major procedure, right, designed to get
• 0:28
  rid of awful pain, give you back your movement.
• 0:30
  But the success of that surgery, it really hinges
• 0:34
  on an almost unbelievable level of accuracy.
• 0:37
  We're talking replacing parts of your shoulder
• 0:39
  joint, the socket, the glenoid, getting that
• 0:41
  angle spot on. You might think, well, how much
• 0:43
  difference can a few degrees make? But, you know,
• 0:45
  in the body, especially a joint like the shoulder.
• 0:47
  Tiny angles, huge consequences. Yeah, think about
• 0:50
  this. Using some of the older methods, a surgeon
• 0:52
  might, without even knowing it, place an implant
• 0:55
  based on a measurement that's off by what the
• 0:57
  paper suggests, over 14 degrees in some cases.
• 0:59
  14 degrees. That's massive in this context. Exactly.
• 1:02
  And that tiny sounding error, it could mean years
• 1:05
  of pain -free life versus going back under the
• 1:08
  knife for another really tough, expensive revision
• 1:11
  surgery. Which nobody wants. Nobody. And this
• 1:13
  isn't just for surgeons, is it? It's really for
• 1:16
  anyone curious about how medicine moves forward,
• 1:19
  how brilliant people are pushing for that accuracy
• 1:21
  that genuinely changes lives. Absolutely. And
• 1:24
  what's really remarkable here, I think, is how
• 1:26
  a seemingly small refinement, just tweaking how
• 1:29
  you measure something, can have these huge ripple
• 1:31
  effects. Our mission today is to unpack one of
• 1:34
  those innovations, a potentially groundbreaking
• 1:38
  way to measure that shoulder socket angle, the
• 1:40
  glenoid version. It's called the ellipse modification
• 1:43
  of the Friedman method. Ellipse modification,
• 1:45
  okay. And this isn't just theory, it's backed
• 1:47
  by solid research, a key paper published in the
• 1:49
  Bone and Joint Journal back in February 2020.
• 1:52
  And a really crucial contributor to this work,
• 1:55
  someone instrumental in actually writing the
• 1:57
  manuscript. So really shaping how the findings
• 1:59
  were presented, ensuring that scientific rigor
• 2:01
  is Professor Mo Imam. Professor Imam? Yes, he's
• 2:05
  based at Ashford in St. Peter's Hospital's NHS
• 2:07
  Trust. And he's, well, he's a highly prolific
• 2:11
  researcher. exceptionally cited. We're talking
• 2:14
  146 publications, over 1 ,100 citations. His
• 2:18
  input here really speaks volumes about the thought
• 2:20
  behind this method. That's some serious expertise.
• 2:23
  It is. And the core idea, the whole point, is
• 2:25
  that getting that precision before surgery in
• 2:28
  the planning stage is just vital for success.
• 2:31
  It directly impacts how well you function afterwards,
• 2:33
  your quality of life. So what does this all mean
• 2:35
  for you listening in? Maybe you're not a surgeon,
• 2:37
  maybe you're not having shoulder surgery, but
• 2:39
  Understanding this, this relentless drive for
• 2:42
  getting things exactly right, it gives you real
• 2:44
  insight into how medicine improves, doesn't it?
• 2:46
  It really does. We're going to get into why these
• 2:49
  tiny angles matter so much and how research like
• 2:52
  this, spearheaded by people like Professor Imam,
• 2:55
  makes a real difference. It's science refining
• 2:58
  how we heal. OK, so let's set the stage properly.
• 3:01
  Let's talk about the stakes in total shoulder
• 3:04
  arthroplasty because they are incredibly high.
• 3:07
  TSA, it's a big deal. You're replacing damaged
• 3:10
  parts. the ball of your upper arm bone, the humerus,
• 3:14
  and that shallow cup, the glenoid, your shoulder
• 3:16
  socket, replacing them with artificial parts.
• 3:19
  Prostetic components, yeah. Think of it like,
• 3:21
  I don't know, building a house inside your shoulder,
• 3:23
  high stakes construction. And just like with
• 3:25
  the house, the foundation, it has to be perfectly
• 3:27
  level if it's off even a bit. Everything else
• 3:30
  is compromised. Walls, roof. Exactly. In TSA,
• 3:33
  that foundation is the glenoid component, this
• 3:36
  artificial socket they put in. And the main goal,
• 3:38
  the absolute driver for the... is getting that
• 3:41
  alignment just right, anatomically correct. That's
• 3:44
  the key to making the implant last, making sure
• 3:46
  your shoulder works properly, hopefully for decades.
• 3:50
  Yeah. Pain -free movement. That's the dream.
• 3:51
  The dream. And connecting this to the bigger
• 3:54
  picture. The reason that foundation, that alignment,
• 3:58
  is so incredibly critical is because glenoid
• 4:00
  loosening. Well, it's the most common reason
• 4:03
  why these anatomical shoulder replacements fail.
• 4:06
  loosening, so the implant comes away from the
• 4:08
  bone. Essentially, yes. And that's the number
• 4:11
  one reason patients end up needing revision surgery,
• 4:13
  a second operation, which is complex, costly,
• 4:18
  painful, just something everyone wants to avoid.
• 4:21
  Definitely. And this is where glenoid version
• 4:22
  comes right into the spotlight. It's basically
• 4:24
  the angle of that socket. Is it tilted too far
• 4:27
  forward? That's anti -version? Yeah. Or too far
• 4:29
  backward? That's retroversion. OK, the tilt.
• 4:32
  Yeah, the tilt. If that angle isn't right, Even
• 4:34
  by just a few degrees, it puts abnormal stress
• 4:37
  on the new implant. Constant, uneven pressure.
• 4:41
  And over time, that stress can cause the component
• 4:43
  to slowly work its way loose. Ah, okay. So that
• 4:47
  slight tilt leads directly to the loosening problem.
• 4:49
  It's a major contributing factor, absolutely.
• 4:51
  And these aren't just, you know, theoretical
• 4:52
  worries from research papers, are they? This
• 4:54
  stuff is real. It hits patients hard. Devastatingly
• 4:58
  real. Yeah. and meticulously documented. There
• 5:01
  are studies, right, like Farron and his team
• 5:04
  back in 2006. They found if the glenoid was put
• 5:06
  in more than five degrees too far back, too retroverted.
• 5:09
  Yeah, more than five degrees retroversion. It
• 5:11
  led to a seven -fold increase in micromotion.
• 5:14
  Seven times. Micro ocean. That sounds small,
• 5:18
  but it's bad, right? It's terrible. Imagine tiny
• 5:21
  little destructive wobbles where the implant
• 5:22
  meets the bone. It destabilizes everything. And
• 5:25
  they also sound a 3 .26 fold increase in stress
• 5:29
  right at that bone cement interface. It's basically
• 5:32
  asking for the implant to fail early. Wow. Okay.
• 5:35
  And then there was Shapiro, 2007. They showed
• 5:38
  that if you don't correct that backward tilt,
• 5:41
  that retroversion, you get eccentric loading.
• 5:43
  Eccentric loading, yeah. Think of it like the
• 5:45
  forest isn't spread evenly across the implant.
• 5:47
  It's constantly pushing it off -center. Like
• 5:49
  leaning on one side of a chair leg constantly.
• 5:51
  Kind of, yeah. And that leads directly to more
• 5:53
  wear on the implant material itself. And again,
• 5:55
  it loosens faster. And even small amounts matter.
• 5:58
  I think Neifler's group showed that. Yes. Neifler
• 6:00
  at all. In 2006, they used cadaver models. Very
• 6:04
  precise. Showed that even tiny steps of retroversion
• 6:07
  just four degrees at a time, caused the humeral
• 6:10
  head, the ball part, to shift backward off center.
• 6:13
  And that led to early loosening in their models.
• 6:15
  So even small errors add up. They absolutely
• 6:17
  do. And maybe most importantly for how a patient
• 6:20
  actually feels and functions, Ye -Yin and colleagues
• 6:23
  in 2005 found a direct link. More retroversion
• 6:27
  in the implant. Significantly lower constant
• 6:29
  scores. The constant score, that measures shoulder
• 6:32
  function, right? Pain, movement. Exactly. It's
• 6:34
  a standard measure. Lower scores mean more pain,
• 6:37
  less movement, basically a shoulder that doesn't
• 6:39
  work well for everyday things, reaching up, getting
• 6:42
  dressed. So we're not just talking about implants
• 6:44
  revival here. We're talking about real day -to
• 6:46
  -day pain and limitation for the patient and
• 6:49
  potentially needing another operation. Precisely.
• 6:52
  And while, sure, other things play a role, patient
• 6:54
  factors like age, sex. the implant design itself,
• 6:58
  getting that glenoid position right, that's a
• 7:00
  huge factor. And crucially, it's something the
• 7:03
  surgeon can control. It's controllable. That's
• 7:05
  the key. It is. It's a critical moment in the
• 7:07
  OR where accurate measurement can literally change
• 7:11
  the patient's future. OK, so the stakes are sky
• 7:14
  high. Getting that angle right is non -negotiable,
• 7:18
  which brings us to the obvious question. How
• 7:21
  do surgeons measure it accurately? You mentioned
• 7:24
  CT scans. Regular x -rays aren't enough. No.
• 7:28
  Conventional x -rays just don't give you the
• 7:30
  detail you need for version. You absolutely need
• 7:33
  a CT scan, computed tomography. It gives you
• 7:35
  those cross -sectional slices. Right. But even
• 7:37
  with CT scans, you said there's no single best
• 7:41
  way everyone agrees on. That's the surprising
• 7:43
  thing. There isn't universal consensus. Different
• 7:45
  techniques exist. But one thing everyone does
• 7:48
  agree on is that the CT scans must be properly
• 7:51
  reformatted. Reformatted? What does that mean?
• 7:53
  It means digitally realigning the images so you're
• 7:55
  looking squarely at the shoulder blade, the scapula,
• 7:57
  and its natural plane. You remove any tilt or
• 8:00
  rotation from how the scan was taken. It's called
• 8:02
  the scapula plane reformat. OK, so you have to
• 8:05
  get the image lined up perfectly first. Absolutely.
• 8:07
  If you don't do that basic step right, any measurement
• 8:10
  you take afterwards is basically garbage. It
• 8:12
  won't be accurate or reproducible. Got it. Foundational
• 8:15
  step. So assuming you've got your properly reformatted
• 8:19
  CT, what's been the main method people use? Historically,
• 8:23
  and still very common because it's relatively
• 8:25
  straightforward, the go -to has been the Friedman
• 8:28
  method, described way back in 1992 by Friedman
• 8:31
  and his colleagues. Friedman method. OK. How
• 8:34
  does that work? It uses a bony landmark called
• 8:37
  the coracoid process. It's this hook -like bone
• 8:40
  sticking off the front of your shoulder blade.
• 8:42
  You find that, and then you measure the glenoid
• 8:44
  version on an axial slice, a cross section that's
• 8:47
  10 millimeters below it. 10 millimeters below
• 8:49
  the coracoid? Seems specific. It is specific.
• 8:52
  Yeah. But, and here's the crucial but, if it's
• 8:55
  so widely used, what's the problem? Why bother
• 8:57
  developing something new like Professor Imam
• 8:59
  and his co -authors did? Yeah, good question.
• 9:01
  What's the catch? The catch is the coracoid itself.
• 9:04
  Ah, the reference point. Exactly. And this is
• 9:07
  where the paper we're looking at gets really
• 9:09
  interesting and where the ellipse method comes
• 9:10
  in. The paper points out quite directly that
• 9:14
  the coracoid process is highly variable. Variable
• 9:18
  how? Like shape? Position? Both. Its shape can
• 9:23
  differ quite a bit between people. and critically,
• 9:26
  its exact position relative to the actual glenoid
• 9:29
  socket isn't consistent. So if your landmark
• 9:31
  moves around... Your measurement starting point
• 9:33
  moves around. It makes it really hard to consistently
• 9:35
  find what everyone agrees is the true mid -glenoid
• 9:39
  level. Mid -glenoid level. Why is heading that
• 9:42
  exact middle point so important? Because the
• 9:44
  glenoid version... that angle we care so much
• 9:46
  about. It actually changes along the face of
• 9:48
  the socket. It's not the same angle top to bottom.
• 9:50
  Really? It varies. Oh, yeah. Typically, the upper
• 9:53
  part of the glenoid is more retroverted, tilted
• 9:56
  back, more than the middle part of the true equator.
• 9:58
  Ah, OK. So if you measure too high or too low
• 10:02
  because your coracoid landmark put you in the
• 10:04
  wrong spot. You get a skewed measurement. You
• 10:06
  might think the version is, say, 10 degrees retroverted,
• 10:09
  but the true version at the middle might only
• 10:11
  be five degrees. or vice versa. You're measuring
• 10:14
  the wrong slice, essentially. And that inconsistency,
• 10:18
  that reliance on a variable landmark, that's
• 10:20
  the Achilles heel of the Friedman method. That's
• 10:23
  exactly what the authors argue. It introduces
• 10:25
  this inherent variability that surgeons have
• 10:28
  just had to kind of deal with. Now there are
• 10:30
  other methods, sure, Randelli's method, Poonin
• 10:33
  -Ting's vault method for... tricky cases. But
• 10:35
  Friedman's has been the most popular. Generally,
• 10:38
  yes, because it seems relatively easy to apply,
• 10:40
  which makes this variability problem even more
• 10:43
  significant. You know, it highlights why a more
• 10:46
  reliable way to find that mid -glenoid level
• 10:48
  was really needed. And what's really cool about
• 10:50
  the ellipse method is where it came from. It
• 10:53
  wasn't just dreamt up in a lab somewhere. It
• 10:55
  came directly out of the clinical trenches, you
• 10:57
  could say. The authors, including Professor Mo
• 11:00
  Imam, who, remember, was key in actually writing
• 11:02
  this paper, articulated the findings they were
• 11:04
  facing, these exact limitations of the Friedman
• 11:06
  method, day in, day out in their own surgical
• 11:09
  practice. They were living the problem. Exactly.
• 11:11
  And especially with newer digital imaging, where
• 11:14
  you can see the axial and sagittal views cross
• 11:17
  -section and side view side by side simultaneously.
• 11:20
  Ah, that must make the inconsistencies more obvious.
• 11:23
  It really does. You start seeing how relying
• 11:25
  on just the coracoid isn't always hitting the
• 11:29
  same spot relative to the glenoid itself. So
• 11:32
  their goal became really clear. Find a way to
• 11:36
  consistently, reliably nail down that mid -glenoid
• 11:38
  level every single time. Not just guess, or hope
• 11:41
  you're in the right place. Precisely. It wasn't
• 11:43
  about totally changing how you measure the angle
• 11:46
  between the lines, but about making sure where
• 11:48
  you draw those lines is standardized and accurate.
• 11:50
  picking the guesswork out. OK, so let's unpack
• 11:52
  it. How does this ellipse method actually work?
• 11:54
  Walk us through the steps. It sounds like a visual
• 11:56
  journey. It is, and it's actually quite elegant
• 11:58
  once you break it down. It's basically five steps
• 12:00
  using those modern CT views. Step one, you start
• 12:04
  with the sagittal view, the side view of the
• 12:05
  shoulder. And on that view, you draw what they
• 12:09
  call a best fit ellipse. Right around the entire
• 12:11
  glenoid fossa, top to bottom. Like putting an
• 12:14
  oval picture frame around the socket from the
• 12:17
  side. That's a perfect analogy. You capture the
• 12:19
  whole vertical extent. That ellipse becomes your
• 12:22
  geometric guide. Step two. Step two. Measure
• 12:25
  the vertical diameter of that ellipse you just
• 12:27
  drew. The total height. Simple measurement. Got
• 12:30
  it. Full height. Step three. Step three. Calculate
• 12:34
  the exact midpoint of that vertical diameter.
• 12:37
  Just divide the height by two. That gives you
• 12:39
  the precise vertical center of the glenoid in
• 12:42
  that side view. The geometric middle, not based
• 12:44
  on a wobbly bone landmark. Exactly. That's the
• 12:46
  key innovation. Use this geometry, not variable
• 12:49
  anatomy, to find the center. Clever. Okay, step
• 12:52
  four. Step four. Now, you use the power of the
• 12:55
  linked views. You find the axial slice, the cross
• 12:58
  -sectional view, that corresponds exactly to
• 13:00
  that calculated midpoint you found on the sagittal
• 13:02
  view. You scroll through the axial slices until
• 13:05
  you hit the one that lines up perfectly with
• 13:06
  your midpoint marker. Ah, so the sagittal view
• 13:09
  guides you to the correct axial slice. Precisely.
• 13:12
  That specific slice is your defined, consistent,
• 13:15
  reproducible mid -glenoid level. Found it. No
• 13:18
  more ambiguity. And step five. Step five. Now
• 13:23
  that you're confidently on the right axial slice,
• 13:26
  you measure the glenoid version, and you use
• 13:29
  the exact same landmarks as the original Friedman
• 13:31
  method. You draw the glenoid line across the
• 13:34
  socket face. You draw the scapular axis line.
• 13:37
  And measure the angle between them. Correct.
• 13:39
  So the angle measurement itself is familiar,
• 13:41
  but the plane you're measuring it on is now defined
• 13:43
  with much, much greater consistency thanks to
• 13:47
  that ellipse. Okay, let me make sure I've got
• 13:48
  this. It's not a totally new angle measurement,
• 13:51
  but a much, much better way of making sure you're
• 13:54
  measuring that angle at the right spot, the true
• 13:57
  middle of the glenoid every single time. You've
• 13:59
  got it. It refines Friedman by fixing its biggest
• 14:01
  weakness. Finding the correct level reliably,
• 14:04
  it gives it a stable geometric foundation. That
• 14:07
  makes a lot of sense. And this leads to a really
• 14:09
  important question. What does this mean, practically,
• 14:12
  for the surgeon in the operating room? Yeah,
• 14:14
  beyond just a better number on a planning sheet.
• 14:17
  Well, finding that true mid -glenoid level accurately,
• 14:21
  it doesn't just help plan. It gives the surgeon
• 14:24
  a much clearer picture of the ideal entry point
• 14:28
  for the central guide wire when they're actually
• 14:30
  putting the glenoid component in. The first pin
• 14:32
  they place to guide the implant. Exactly. The
• 14:35
  paper calls it intraoperative corroboration.
• 14:38
  It means the surgeon has a better map during
• 14:40
  the surgery itself. They could be more confident
• 14:43
  they're aiming for the right spot, the center
• 14:45
  they identified in planning. So it connects the
• 14:47
  plan to the action much more reliably. That's
• 14:50
  the idea. Potentially fewer errors in placement,
• 14:54
  better positioning of the actual implant, it
• 14:56
  builds that bridge between planning and doing.
• 14:58
  Okay, so the message sounds elegant, logical.
• 15:01
  Very clever, actually. But how did they prove
• 15:03
  it works? How did they test it rigorously? Right.
• 15:05
  The study design. This was a retrospective study,
• 15:07
  meaning they went back and looked at data that
• 15:09
  already existed. Sedation scans already done.
• 15:11
  Exactly. They pulled up 100 CT scans from 100
• 15:14
  different patients who were all getting their
• 15:16
  first primary shoulder replacement. So a good
• 15:18
  size group. An irrelevant group, the actual patients
• 15:21
  who need this. Definitely. It was a mix, 58 women,
• 15:24
  42 men, average age around 65, 66, but ranging
• 15:28
  all the way from 19 to 93. So it captured that
• 15:31
  typical TSA demographic pretty well. And the
• 15:33
  stands themselves, they made sure they were suitable?
• 15:36
  Oh, absolutely crucial. They used a standard
• 15:38
  protocol, thin, one millimeter slices through
• 15:40
  the glenoid. And importantly, they ensured all
• 15:44
  scans were properly reformatted into that scapula
• 15:47
  plane we talked about. Giving the alignment right
• 15:48
  from the start. Essential. They also had clear
• 15:51
  rules for who was in and who was out inclusion
• 15:53
  -exclusion criteria. Had to be adults, first
• 15:57
  -time TSA, needed a good quality CT scan available
• 16:00
  digitally. And who did they exclude? People who'd
• 16:03
  already had shoulder replacements or surgery
• 16:05
  on their scapula, anyone whose bones weren't
• 16:07
  fully mature, or if the CT scan was just technically
• 16:10
  inadequate, not properly reformatted, didn't
• 16:13
  show the whole shoulder blade, that sort of thing.
• 16:15
  So they were careful to compare like with like,
• 16:18
  making sure the data was clean. Very careful.
• 16:20
  That's vital for results you can trust. Okay,
• 16:22
  so they had the scans. Who did the measuring
• 16:25
  and how? Two senior observers did the measurements.
• 16:28
  The paper calls them SJ and SH. Experienced clinicians
• 16:32
  who know how to read these scans. And they both
• 16:35
  measured glenoid version using the classic Friedman
• 16:37
  method and the new ellipse method on all 100
• 16:40
  scans. So a direct comparison on the same patients.
• 16:43
  Both methods, both observers, okay. But they
• 16:45
  added another layer of rigor, observer 1, SJ.
• 16:49
  They measured Friedman once, but they measured
• 16:52
  using the ellipse method twice. Twice. Why? With
• 16:55
  a one -month gap between the two measurements.
• 16:57
  This was to test intra -rater reliability. How
• 17:00
  consistent is the same person using the ellipse
• 17:02
  method over time? Do they get the same result
• 17:05
  if they measure again later? Yeah, checking for
• 17:07
  drift or inconsistency in one person's measurements.
• 17:10
  Smart. Very smart. And observer two, SH, measured
• 17:13
  the ellipse method once. So this whole setup
• 17:15
  allowed them to look at everything. Likewise.
• 17:16
  Well, comparing ellipse versus Friedman directly.
• 17:19
  Comparing observer 1 versus observer 2 using
• 17:23
  the ellipse method that's iterator reliability.
• 17:26
  Consistency between different people. And comparing
• 17:29
  observer one's first ellipse measurement to their
• 17:31
  second Heller, that's the inter -rater reliability
• 17:33
  we just talked about. Wow, okay. That's thorough.
• 17:36
  Covering all the bases for reliability. Extremely
• 17:39
  thorough. It's how you really build confidence
• 17:41
  in a new measurement technique. So they had all
• 17:43
  these measurements. How did they analyze the
• 17:45
  numbers? What stats did they use? They used a
• 17:47
  solid statistical toolkit to really dig into
• 17:50
  what the numbers meant. First, a paired t -test.
• 17:53
  This basically checks if the average difference
• 17:56
  between two sets of measurements like Ellipse
• 17:58
  versus Friedman or Observer 1 versus Observer
• 18:00
  2 is statistically significant. Is the difference
• 18:02
  likely real or just random chance? Okay, average
• 18:05
  difference. Then, critically, they use bland
• 18:07
  Altman plots. These are brilliant for looking
• 18:10
  at agreement between methods. Instead of just
• 18:12
  averages, it plots the difference between two
• 18:15
  measurements for each patient against the average
• 18:17
  of those two measurements. So you see the spread.
• 18:19
  How much individual measurements might disagree?
• 18:21
  Exactly. It visually shows you the range of disagreement.
• 18:25
  Are the methods generally close across the board,
• 18:27
  or do they sometimes differ wildly? You look
• 18:30
  for the limits of agreement on the plot. Narrow
• 18:33
  limits mean good agreement. Wide limits? Not
• 18:36
  so much. Okay, Bland -Altman for agreement. What
• 18:39
  else? The Interclass Correlation Coefficient,
• 18:41
  or ICC. This gives you a single number between
• 18:44
  0 and 1 that quantifies reliability or consistency.
• 18:48
  Higher is better. In this study, they defined
• 18:50
  0 .90 or above as excellent. ICC for reliability.
• 18:54
  Got it. And finally, the repeatability coefficient.
• 18:57
  This gives you a practical number. It tells you
• 18:59
  how much difference you'd expect to see between
• 19:01
  two measurements made by the same person using
• 19:03
  the same method. 95 % of the time. A smaller
• 19:06
  number means better repeatability, higher precision.
• 19:09
  So t -test for significance, bland Altman for
• 19:11
  agreement spread, ICC for overall reliability,
• 19:15
  and repeatability for single user precision.
• 19:17
  That's comprehensive. It really is. And one more
• 19:19
  crucial thing. They defined beforehand what they
• 19:23
  considered clinically significant. Ah, not just
• 19:25
  statistically significant, but actually mattering
• 19:28
  in practice. Exactly. For this study, they set
• 19:30
  that threshold at a difference greater than plus
• 19:33
  or minus five degrees between methods. Five degrees.
• 19:36
  Why five? That's generally considered a threshold
• 19:38
  in orthopedics, where the difference is large
• 19:41
  enough that it might actually change the surgeon's
• 19:43
  plan, affect implant choice, positioning strategy
• 19:46
  and potentially impact the patient's outcome.
• 19:48
  Okay. So they had a clear line in the sand for
• 19:51
  what really matters clinically. Yes. It allows
• 19:54
  them to interpret the statistical results in
• 19:56
  a really practical, meaningful way for surgeons
• 19:59
  and patients. All right. The moment of truth.
• 20:01
  They did all this careful work. What did the
• 20:04
  numbers actually show? Let's start with the average
• 20:06
  differences. Friedman versus ellipse. OK, so
• 20:09
  the average glenoid version measured by the traditional
• 20:12
  Friedman method was maggots of 3 .11 degrees.
• 20:16
  That negative means retroversion tilted slightly
• 20:18
  back. OK, average Friedman minus 3 .11. The average
• 20:22
  measured by the ellipse method using observer
• 20:24
  1's data was minus the 1 .95 degrees retroversion.
• 20:28
  Minus 1 .95, so slightly less retroverted on
• 20:31
  average with the ellipse method. On average,
• 20:34
  yes. The mean difference between Ellipse and
• 20:37
  Friedman was 1 .16 degrees. And statistically,
• 20:40
  this difference was significant. The p -value
• 20:42
  was 0 .002. OK, 1 .16 degrees difference on average,
• 20:47
  and statistically significant. But you hinted
• 20:50
  earlier that the average isn't the full story,
• 20:52
  right? Not even close. This is exactly where
• 20:55
  it gets really interesting and why those bland
• 20:57
  Altman plots are so vital. That 1 .16 degree
• 21:01
  average difference. Honestly, it might sound
• 21:03
  pretty small. Maybe not clinically huge on its
• 21:05
  own. Yeah, you might think, okay, about a degree,
• 21:07
  maybe not a disaster. Right. But that average
• 21:10
  hides the spread. It tells you nothing about
• 21:12
  how much the methods might disagree for an individual
• 21:14
  patient. And that's where the blind Altman plot
• 21:16
  comes in, comparing Ellipse versus Friedman.
• 21:19
  Exactly. And the results there were... Well,
• 21:21
  quite startling, actually. The 95 % limits of
• 21:23
  agreement were huge. They ranged from minus 6
• 21:26
  .15 degrees all the way up to plus 8 .46 degrees.
• 21:30
  Oh, minus 6 to plus 8 .5. What does that mean,
• 21:33
  practically? It means. OK, let's say the ellipse
• 21:36
  method measures a patient's glenoid at exactly
• 21:39
  0 degrees version. The Bland -Altman plot tells
• 21:42
  us that for that same patient, the Friedman method
• 21:45
  could, 95 % of the time, report value anywhere
• 21:48
  between meganis 6 .15 degrees and plus 8 .46
• 21:52
  degrees. That's a potential range of almost 15
• 21:55
  degrees for the same shoulder, depending only
• 21:57
  on which method you used. Precisely. That represents,
• 22:01
  in the author's words, and by any clinical standard,
• 22:04
  poor... agreement. The two methods might be close
• 22:07
  on average, but for any given patient, they could
• 22:09
  give you wildly different numbers. And that could
• 22:11
  lead to totally different surgical decisions
• 22:13
  for the same person. Absolutely. And remember
• 22:15
  that five degree threshold for clinical significance.
• 22:17
  Yeah. The study found that in 16 % of the patients,
• 22:20
  16 out of 100, the difference between the ellipse
• 22:23
  measurement and the Friedman measurement was
• 22:25
  more than five degrees. 16%. Nearly one in five
• 22:28
  patients had a clinically significant difference
• 22:30
  depending on the method used. Yes. For almost
• 22:33
  a fifth of these patients, the choice of measurement
• 22:36
  method could have directly led to a different
• 22:38
  surgical plan, potentially impacting their outcome.
• 22:41
  It really drives home how much finding a consistent
• 22:43
  method matters. Wow. Okay, so poor agreement
• 22:46
  between the methods, but how did the ellipse
• 22:48
  method perform on its own? You mentioned they
• 22:50
  tested its internal consistency. Ah, now that's
• 22:53
  where the story flips. In stark contrast to the
• 22:56
  poor agreement with Friedman, the ellipse method
• 22:58
  showed excellent reliability and reproducibility,
• 23:01
  both between observers and within the same observer.
• 23:04
  Okay, let's break that down. First, between the
• 23:06
  two observers, interrater reliability. Right.
• 23:08
  Comparing observer 1 and observer 2, using the
• 23:11
  ellipse method, the mean difference between them
• 23:13
  was tiny. It's just... 0 .063 degrees, not statistically
• 23:18
  significant at all. Basically, they got the same
• 23:20
  average result. Good start. And the bland Altman
• 23:22
  limits for ellipse between observers? Were they
• 23:25
  wide, like before? Remarkably narrow. The 95
• 23:28
  % limits of agreement were only 1 .98 to plus
• 23:32
  2 .0 degrees. Minus 2 to plus 2. So 95 % of the
• 23:36
  time, two different surgeons using the ellipse
• 23:40
  method would get a result within about two degrees
• 23:42
  of each other for the same patient. Exactly.
• 23:44
  That's excellent agreement. Much, much better
• 23:46
  than the nearly 15 degree range we saw when comparing
• 23:49
  with Friedman. Okay. And the ICC value for that?
• 23:52
  The interclass correlation coefficient for iterator
• 23:54
  agreement was 0 .993. 0 .993. That's way above
• 23:59
  the 0 .90 threshold for excellent. Absolutely.
• 24:02
  It indicates excellent reproducibility between
• 24:04
  different users. Fantastic. Okay, now what about
• 24:06
  iterator? The same observer measuring twice a
• 24:09
  month apart. Even better, actually. Observer
• 24:11
  1, measuring the same scans twice with the ellipse
• 24:14
  method. The mean difference was minuscule. 9
• 24:16
  is 0 .8 or 41 degrees. Again, not statistically
• 24:19
  significant. Almost zero difference on average.
• 24:22
  Pretty much. And then 95 % limits of agreement.
• 24:24
  Even tighter. From 9 is .98 degrees to plus .90
• 24:27
  degrees. Less than one degree variation 95 %
• 24:31
  of the time if the same person measures again.
• 24:33
  Incredible, isn't it? That's superb precision.
• 24:36
  And the repeatability coefficient confirmed this.
• 24:38
  It was 0 .94 degrees, meaning 95 % of repeated
• 24:42
  measurements by the same person will differ by
• 24:45
  less than one degree. Wow. And the ICC for IntraRater?
• 24:48
  Almost perfect. 0 .999. OK, so to summarize,
• 24:52
  ellipse versus Friedman. Poor agreement, potentially
• 24:55
  large clinically significant differences, but
• 24:57
  ellipse versus ellipse, whether between users
• 24:59
  or by the same user over time. Excellent agreement,
• 25:03
  highly reproducible, very precise. That's the
• 25:05
  core finding in a nutshell. The ellipse method
• 25:08
  demonstrated a level of reliability that the
• 25:10
  traditional Friedman method simply couldn't match
• 25:12
  in this study. So this really raises the question,
• 25:14
  doesn't it? What could this level of proven precision
• 25:17
  mean for actual shoulder surgery moving forward?
• 25:21
  It suggests we now have a tool, a method, that
• 25:23
  can provide a much more trustworthy and consistent
• 25:26
  number for that critical glenoid version angle.
• 25:29
  You can standardize planning, reduce variability
• 25:31
  between surgeons, and give everyone more confidence
• 25:34
  in that initial assessment. And that confidence
• 25:36
  should, hopefully, translate into better surgical
• 25:39
  execution and, ultimately, better patient outcomes.
• 25:43
  Okay, let's zoom out a bit. Let's talk broader
• 25:45
  context and really hammer home the so what of
• 25:48
  all this The main message from the paper then
• 25:51
  isn't just here's a new method It's that the
• 25:54
  real breakthrough is having a reliable way to
• 25:57
  find that mid glenoid level, right? That's what
• 25:59
  the authors including professor Imam in his writing
• 26:02
  really emphasize They call consistently defining
• 26:05
  that mid glenoid level the most important step
• 26:08
  for getting accurate 2d CT measurements of version
• 26:10
  and it's so important because Other ways of trying
• 26:13
  to find it, like just counting slices down from
• 26:15
  the top or something. Yeah. They just weren't
• 26:16
  reliable enough. Exactly. The paper cites another
• 26:19
  study, Vandebunt et al., from 2015, that tried
• 26:22
  a slice counting method. Their ICC for reliability,
• 26:25
  only .70. .70 versus the ellipse method's .993
• 26:31
  between observers. That's a night and day difference
• 26:34
  in reliability. It really is. It shows the ellipse
• 26:37
  method isn't just a small tweak. It's a significant
• 26:40
  leap in consistency. It takes the ambiguity out
• 26:44
  of where you measure, ensuring everyone's looking
• 26:46
  at the same anatomically relevant spot. Right.
• 26:49
  Getting everyone on the same page measurement
• 26:51
  -wise. And this touches on something else. The
• 26:53
  paper highlights the importance of using the
• 26:55
  right statistical tools and using them together.
• 26:58
  The Bland -Altman plots and the ICCs. Yes. Professor
• 27:01
  Imam and his co -authors stress this point in
• 27:03
  the manuscript. The Bland -Altman shows you the
• 27:06
  side of the differences, the practical agreement.
• 27:08
  The ICC gives you that overall reliability number.
• 27:11
  Using both gives you a much better picture of
• 27:14
  the true reliability. So you see not just if
• 27:17
  they're correlated, but how well they actually
• 27:19
  agree within limits that matter clinically. Precisely.
• 27:22
  It's about robust science. Not just finding a
• 27:24
  difference, but understanding its magnitude and
• 27:26
  consistency. That kind of methodological rigor
• 27:29
  is just as important as the clinical idea itself.
• 27:33
  Absolutely. And you mentioned this study's reliability
• 27:35
  for the ellipse method was better than what other
• 27:37
  studies found when comparing Friedman to other
• 27:39
  methods like Randellis or the vault method. That's
• 27:42
  right. Comparing the ICCs reported here for the
• 27:45
  ellipse method, those high 0 .99 values, with
• 27:49
  ICCs reported in studies by Rouleau et al. comparing
• 27:52
  Friedman and Randalli, or Mezzimora at all, comparing
• 27:54
  Friedman and Valt, the ellipse method demonstrated
• 27:57
  superior consistency in this particular investigation.
• 28:00
  It really seems to set a new benchmark for 2D
• 28:03
  measurement liability. OK. But what about 3D?
• 28:06
  We hear a lot about 3D planning and surgery now.
• 28:10
  Does this 2D ellipse method still matter in the
• 28:13
  age of 3D? That's a really important question.
• 28:15
  And the paper acknowledges 3D reconstruction.
• 28:17
  For really complex cases, severe deformities,
• 28:20
  3D planning is often seen as the benchmark. It
• 28:21
  gives you that full volumetric picture. But it
• 28:24
  has limitations for routine, everyday use right
• 28:27
  now. It's generally expensive, needs special
• 28:29
  software, powerful computers. It requires significant
• 28:32
  training and expertise to use properly. OK. Cost
• 28:35
  and training hurdles. And even with 3D, there's
• 28:38
  still user variability. The surgeon still has
• 28:41
  to define reference points and axes on the 3D
• 28:44
  model, which can introduce inconsistencies. Ah,
• 28:47
  so 3D isn't automatically perfect either. Not
• 28:49
  necessarily. And maybe the biggest point is the
• 28:52
  direct link between using 3D planning and actually
• 28:55
  achieving better long -term clinical outcomes
• 28:58
  for patients, like lower revision rates, better
• 29:00
  function scores, years later that link. still
• 29:03
  needs more robust, independent proof across the
• 29:06
  board. So the jury's still out on whether 3D
• 29:08
  consistently leads to better results yet? The
• 29:11
  evidence is still building. And until we have
• 29:13
  that definitive proof, and maybe until 3D becomes
• 29:15
  more accessible and standardized, having a really
• 29:18
  accurate, reliable 2D measurement, a precise
• 29:21
  numerical angle remains incredibly important.
• 29:23
  Why? Why keep focusing on the 2D number? For
• 29:26
  clear communication between surgeons. For comparing
• 29:28
  results in research studies effectively. for
• 29:31
  tracking outcomes over time. You need that consistent,
• 29:34
  quantifiable angle. And that's exactly what the
• 29:36
  highly reliable ellipse method provides. It fills
• 29:39
  a crucial need, even as 3D evolves. Okay, that
• 29:42
  makes sense. So, let's boil it down. What's the
• 29:45
  absolute core take -home message from Professor
• 29:48
  Imam and the other authors of this paper? They
• 29:51
  really drive home three key points. One... Getting
• 29:55
  that preoperative glenoid version measurement
• 29:57
  right is absolutely crucial for planning and
• 30:00
  executing a successful TSA. It's foundational.
• 30:03
  Non -negotiable first step. Two, if you have
• 30:06
  a good quality CT scan that includes the whole
• 30:08
  scapula and is properly reformatted in the scapular
• 30:12
  plane, the ellipse method is a highly reproducible
• 30:15
  and reliable way to find the true mid -glenoid
• 30:18
  level. That's its unique strength. Reliable identification
• 30:20
  of the mid -glenoid. And three, Measuring version
• 30:23
  at that specific mid -glenoid level refines the
• 30:26
  old Friedman method. It gives you an accurate
• 30:28
  2D angle measurement, plus it crucially shows
• 30:30
  the surgeon the ideal spot for that central guide
• 30:33
  wire during surgery. It offers that valuable
• 30:35
  intraoperative corroboration. So better planning
• 30:37
  and more confidence during the actual operation.
• 30:40
  That's the promise. Precision translating into
• 30:42
  practice. Now good science always acknowledges
• 30:44
  its limits, and this study is no exception. The
• 30:47
  authors are upfront about the limitations. OK,
• 30:49
  what are the main ones we should keep in mind?
• 30:51
  Well, first and foremost, this all hinges on
• 30:54
  having that CT scan correctly oriented and reformatted
• 30:58
  in the scapular plane to begin with. Right, the
• 31:00
  foundational step we talked about. Exactly. If
• 31:03
  that basic requirement isn't met, if the initial
• 31:06
  imaging isn't done right, then even the ellipse
• 31:09
  method will give you an erroneous measurement.
• 31:12
  It's not magic. It relies on good data in. So
• 31:15
  the quality of the initial scan is paramount.
• 31:17
  That could be a practical hurdle for widespread
• 31:20
  use if imaging protocols aren't standardized
• 31:23
  everywhere. It certainly could be. It highlights
• 31:26
  the need for consistent, high -quality imaging
• 31:28
  standards if we want to leverage methods like
• 31:30
  this effectively across different hospitals and
• 31:33
  clinics. Okay. What else? The study was retrospective.
• 31:35
  They looked back at past data. Right. So they
• 31:38
  couldn't directly compare patient outcomes. Correct.
• 31:40
  Because they were looking at scans that had already
• 31:42
  been done, they couldn't say, OK, group A was
• 31:44
  planned with ellipse, group B with Friedman,
• 31:46
  and group A had fewer revisions five years later.
• 31:49
  They could only assess the reliability of the
• 31:51
  measurements themselves. But they do argue that
• 31:53
  better planning generally leads to better outcomes,
• 31:56
  right? Yes. There's a strong consensus in orthopedics
• 31:59
  that accurate planning improves results. But
• 32:02
  to directly prove that the ellipse method causes
• 32:05
  better long -term outcomes would require a future
• 32:08
  prospective study. Following patients forward
• 32:10
  in time. Exactly. Comparing groups planned with
• 32:13
  different methods and tracking their function,
• 32:16
  pain scores, implant survival rates over several
• 32:18
  years. That's the next logical step in research.
• 32:21
  Got it. Any other limitations or maybe strengths
• 32:24
  to balance those? A key strength was that they
• 32:27
  used consecutive CT scans. They didn't cherry
• 32:30
  pick easy or hard cases. They took the scans
• 32:32
  as they came. representing the real mix of patients'
• 32:35
  assurgencies. That makes the findings on measurement
• 32:38
  reliability much more generalizable. So it reflects
• 32:41
  real -world practice better. Yes. And they also
• 32:44
  noted that the ellipse method seemed usable even
• 32:46
  when the glenoid was significantly deformed,
• 32:49
  which adds to its potential versatility. Okay.
• 32:52
  So, thinking about all this... What does it mean
• 32:55
  for you, the listener, looking ahead? We've seen
• 32:57
  how this, well, quite elegant refinement in measurement,
• 33:01
  driven by clinical need and rigorous research
• 33:03
  involving experts like Professor Imam, can boost
• 33:06
  precision. Where does that lead? Well, you start
• 33:09
  to think about the knock -on effects, don't you?
• 33:11
  Like how it might change training. Could this
• 33:14
  become the standard way orthopedic residents
• 33:16
  are taught to assess glenoid version, making
• 33:19
  this level of precision just standard practice
• 33:22
  from day one? It certainly could influence curriculum
• 33:24
  development, yeah. Standardizing the approach.
• 33:26
  Then what about technology? Could this method
• 33:29
  be built into surgical planning software? Maybe
• 33:32
  even automated with AI to measure a version super
• 33:34
  consistently, taking human variability out even
• 33:37
  further? That's a definite possibility. Integrating
• 33:39
  validated, reliable methods into software is
• 33:42
  a clear pathway for innovation, reducing potential
• 33:45
  for error. And the big picture. If this leads
• 33:48
  to more accurate surgery, fewer implants failing
• 33:51
  early, what about the savings for health care
• 33:53
  systems, fewer expensive revision surgeries?
• 33:56
  That's a potential long -term benefit. Absolutely.
• 33:58
  Better outcomes often mean lower overall costs
• 34:01
  down the line. So it leaves you wondering, doesn't
• 34:03
  it? What's the next piece of research you'd want
• 34:06
  to see? Is it that direct link to patient outcomes?
• 34:10
  Studies quantifying fewer revisions, faster recovery,
• 34:13
  better long -term comfort, specifically because
• 34:15
  the ellipse method was used. Those outcome studies
• 34:18
  are definitely key. Connecting the improved measurement
• 34:20
  directly to improve patient lives is the ultimate
• 34:23
  goal. It really shows how this constant push
• 34:26
  for precision, even in something as specific
• 34:28
  as measuring an angle, is what drives medicine
• 34:31
  forward, aiming for that perfect outcome one
• 34:34
  step at a time. Well, that was a truly deep dive,
• 34:36
  wasn't it, into orthopedic precision, medical
• 34:38
  innovation, how research like the paper co -authored
• 34:42
  and articulated by Professor Mo Imam can take
• 34:44
  an existing method and just refine it, make it
• 34:46
  so much more reliable, potentially leading to
• 34:49
  really significant improvements for patients.
• 34:50
  Yeah, we went from understanding why glenoid
• 34:52
  loosening is such a critical issue in shoulder
• 34:55
  replacements through the problems with older
• 34:57
  measurement methods, right, to this elegant ellipse
• 35:00
  solution and what it could mean for accuracy.
• 35:02
  It's a fantastic example of how focusing on precision,
• 35:06
  even on what seems like a small detail, can have
• 35:08
  these huge downstream effects in medicine. Absolutely.
• 35:12
  That relentless pursuit of getting it right,
• 35:15
  of building a better foundation through accurate
• 35:16
  measurement, that's how patient care truly advances.
• 35:20
  one reliable step at a time. We hope you found
• 35:22
  that insightful. Join us next time for another
• 35:24
  deep dive, where we'll unpack more sources and
• 35:26
  extract those crucial insights for you. Until
• 35:29
  then, stay curious, stay informed.