EPISODES

two men at table with microphones on table

Advancing Diagnostic Testing and AI Integration: A Conversation with Dr. Michael Laposata

MP3

  •  

    In this episode of The Pulse, Dr. Jochen Reiser, UTMB President and CEO of the UTMB Health System, sits down with Dr. Michael Laposata, Chair of Pathology, for a conversation about his inspiring journey into medicine and the defining experiences that shaped his career. Dr. Laposata discusses his pivotal work in the lab, highlighting how his team is advancing diagnostic processes and creating innovative approaches to patient care. He shares his vision for the future of pathology, emphasizing the importance of integrating laboratory medicine into the broader health care landscape to improve outcomes. Reflecting on his past, Dr. Laposata also underscores the value of mentorship and fostering a culture of discovery and collaboration in advancing medicine.

     

  • 1
    00:00:00,840 --> 00:00:09,600
    [Auto-generated transcript. Edits may have been applied for clarity.]
    Welcome to The Pulse: A UTMB health podcast, your gateway to the latest insights, ideas and innovations shaping our community and beyond.

    2
    00:00:10,230 --> 00:00:16,530
    Stay tuned for engaging discussions, compelling stories, and the information you need to stay informed and connected.

    3
    00:00:16,830 --> 00:00:24,650
    Right here on The Pulse. Hello and welcome back to the Pulse podcast.

    4
    00:00:25,190 --> 00:00:30,800
    I'm Dr. Johan Reiser, president of UTMB and CEO of the UTMB Health System.

    5
    00:00:31,310 --> 00:00:36,490
    On today's episode of The Pulse, I'm excited to welcome Dr. Michael Laposata,

    6
    00:00:36,740 --> 00:00:40,250
    professor and chairman of the Department of Pathology at UTMB.

    7
    00:00:41,060 --> 00:00:49,760
    Dr. Laposata is a nationally recognized leader in his field, making significant contributions to the way we approach diagnostic medicine.

    8
    00:00:50,360 --> 00:00:54,770
    His work is shaping the future of health care here at UTMB and beyond.

    9
    00:00:55,310 --> 00:00:59,750
    Dr. La Laposata, thank you for joining us today. It's great to have you here.

    10
    00:00:59,900 --> 00:01:07,760
    Thank you. Wonderful for me. Please tell me a little bit about your background and what led you to UTMB.

    11
    00:01:08,510 --> 00:01:12,470
    Sure. I've been here 11 years now. A very exciting 11 years.

    12
    00:01:12,830 --> 00:01:17,780
    I started out in the Johns Hopkins program for MD/PhD students.

    13
    00:01:18,260 --> 00:01:21,560
    And during that time I became interested in blood clotting.

    14
    00:01:22,040 --> 00:01:28,070
    So I went to the chairman of hematology at Hopkins and said, how do I become like you?

    15
    00:01:28,790 --> 00:01:32,149
    And he said, well, first you're going to have to be a pathologist.

    16
    00:01:32,150 --> 00:01:35,840
    And I said, how will that work if I'm interested in blood clotting?

    17
    00:01:36,350 --> 00:01:43,160
    He said, well, hematology and oncology emerged. And the hematology part is malignant hematology.

    18
    00:01:43,490 --> 00:01:47,820
    So if you want to study bleeding and clotting, you have to go through pathology.

    19
    00:01:47,840 --> 00:01:52,520
    That turned out to be a good thing because I learned all the tests.

    20
    00:01:52,790 --> 00:01:54,860
    And I also learned how to take care of patients.

    21
    00:01:55,250 --> 00:02:01,670
    So I had a foot in the lab for diagnostics, and I had a foot in front of the patients understanding what was needed.

    22
    00:02:02,300 --> 00:02:09,260
    So I took jobs in my field, running coagulation labs and seeing patients first at Penn for four years.

    23
    00:02:09,500 --> 00:02:14,510
    Harvard for 19 at the mass General started their first modern laboratory.

    24
    00:02:14,780 --> 00:02:18,950
    And then I spent six and a half years practicing for this job by going to Vanderbilt.

    25
    00:02:19,520 --> 00:02:24,830
    Wonderful, wonderful. Can you share more about your current roles at UTMB?

    26
    00:02:24,950 --> 00:02:28,640
    I hear you're balancing three different positions, which is quite impressive.

    27
    00:02:28,820 --> 00:02:34,490
    Thank you. Well, it's all evolved over time because I was interested in coagulation.

    28
    00:02:34,490 --> 00:02:40,100
    One of the first things I noticed is that coagulation results had to be explained to the doctors.

    29
    00:02:40,280 --> 00:02:43,670
    We had to tell them what the test meant and what they should do next.

    30
    00:02:44,060 --> 00:02:50,030
    So ultimately that evolved, and it started out as something that wasn't very scalable,

    31
    00:02:50,180 --> 00:02:58,219
    where we could do 10 or 15 cases in an hour or two, but we realized that the need was much bigger and it was needed in auto-immunity.

    32
    00:02:58,220 --> 00:03:01,310
    It was needed in all kinds of fields of medicine.

    33
    00:03:01,670 --> 00:03:07,670
    So technology arrived, and then with artificial intelligence and large language models,

    34
    00:03:07,970 --> 00:03:14,390
    we realized through the COVID pandemic that we could interpret as many as 2000 cases of anything in a day.

    35
    00:03:14,990 --> 00:03:18,800
    So it was during that time that we learned how to move forward.

    36
    00:03:19,040 --> 00:03:27,080
    So the role of being chairman of the department evolved into a role of being head of a diagnostic center, as we call it.

    37
    00:03:27,560 --> 00:03:30,860
    And, uh... and I'm also head of the MD PhD program.

    38
    00:03:31,040 --> 00:03:34,280
    Wow. It's just incredible what you all covering.

    39
    00:03:34,730 --> 00:03:40,610
    So you referred to it... One of the roles is director of the Sealy Diagnostic Center at UTMB.

    40
    00:03:40,910 --> 00:03:45,660
    How does the center work and is it still in preparation or open for business.

    41
    00:03:45,680 --> 00:03:50,330
    So we're doing work right now mostly for UTMB patients.

    42
    00:03:50,750 --> 00:03:59,150
    So for example um, if a patient comes in and is found to have anemia, we then can respond to a consult.

    43
    00:03:59,600 --> 00:04:06,860
    Um, and we look at everything in the medical record to determine what kind of anemia it is,

    44
    00:04:06,860 --> 00:04:13,010
    because that will tell us how to treat it so we can diagnose simple anemia like iron deficiency.

    45
    00:04:13,160 --> 00:04:21,110
    Or I just tested the system today to see if it can find a rare coagulation disorder called Glanzmann thrombasthenia.

    46
    00:04:21,380 --> 00:04:24,980
    And I told it to platelet aggregation studies and a little about the patient.

    47
    00:04:25,250 --> 00:04:31,850
    And it gave me this language model with the knowledge base that we've created in the diagnostic center,

    48
    00:04:32,000 --> 00:04:38,510
    what the diagnosis was, what we should do for the next set of tests, and what the genetic alterations might be.

    49
    00:04:38,930 --> 00:04:46,160
    So running this diagnostic center has allowed us to take care of UTMB patients for a variety of areas.

    50
    00:04:46,190 --> 00:04:51,350
    Pharmacogenomics is another one. You don't respond to medicines well for all those we can be paid,

    51
    00:04:51,740 --> 00:04:57,379
    but we really think that this is a market that virtually all of us need, so it could go well beyond UTMB.

    52
    00:04:57,380 --> 00:04:59,000
    Be very interesting.

    53
    00:04:59,420 --> 00:05:08,940
    Is the Sealy Diagnostic Center using AI already, and in what way? Can you give a few examples of how AI can help doctors and their patients?

    54
    00:05:08,960 --> 00:05:14,660
    Yeah. First, let me tell you a little bit about AI when it comes to laboratory tests exclusively.

    55
    00:05:14,960 --> 00:05:20,960
    So um, in the example that I just gave, uh, one of our pathologists, Chris Zauner.

    56
    00:05:21,650 --> 00:05:29,540
    Created more than 50 different large language models that cover about 80% of the known diseases.

    57
    00:05:30,110 --> 00:05:34,170
    And so Glanzmann thrombasthenia has its pages and it....

    58
    00:05:34,180 --> 00:05:40,910
    And, what we do is to use artificial intelligence and say... interrogate our large language models.

    59
    00:05:41,060 --> 00:05:44,210
    We might subsequently ask it to check the internet,

    60
    00:05:44,570 --> 00:05:50,000
    but the large language model then picks the verbiage and makes it understandable

    61
    00:05:50,000 --> 00:05:56,120
    to the doctor, so that we can then present the information to the doctor.

    62
    00:05:56,420 --> 00:06:01,399
    So, we are certainly using it now for UTMB patients.

    63
    00:06:01,400 --> 00:06:08,270
    And, and uh, we will go forward. Artificial intelligence is the part where it puts all of the facts together.

    64
    00:06:08,720 --> 00:06:14,090
    Now that's only half the story. The other half of artificial intelligence is for image studies.

    65
    00:06:14,360 --> 00:06:20,420
    So, we know in radiology that artificial intelligence can be used to look at an MRI.

    66
    00:06:20,480 --> 00:06:24,740
    And, it finds what the anomaly is. So why not a biopsy.

    67
    00:06:24,860 --> 00:06:26,870
    Why can't we use it to find a tumor.

    68
    00:06:27,170 --> 00:06:35,480
    And, just this past Tuesday, we turned on artificial intelligence, one of two academic medical centers in the United States today.

    69
    00:06:35,840 --> 00:06:42,260
    And, it was so impressive. Everybody stood up like a rocket launch when Harsh Thaker, head of anatomic pathology,

    70
    00:06:42,500 --> 00:06:49,160
    push the button and suddenly all the tumor lit up in red and it was not visible before.

    71
    00:06:49,520 --> 00:06:58,940
    So really with both anatomic path and clinical path, we're picking up lots of things that we never, never would have because of AI.

    72
    00:06:59,300 --> 00:07:04,300
    Wonderful. I actually shared that moment on the social media platform.

    73
    00:07:04,310 --> 00:07:06,470
    I was so excited about it. Yes.

    74
    00:07:06,950 --> 00:07:15,290
    So it seems like our patients are already being able to participate in all these new technologies in your pathology department.

    75
    00:07:15,770 --> 00:07:21,420
    And, and are they covered these tests that you're mentioning, uh, these new technologies by insurance.

    76
    00:07:21,440 --> 00:07:30,350
    Yes. So certainly the prostate biopsy was covered because they needed a prostate biopsy interpretation, and we did it faster and more reliably.

    77
    00:07:31,130 --> 00:07:36,110
    Uh, certainly the anemia case was covered because the patient was anemic and needed an interpretation.

    78
    00:07:36,500 --> 00:07:39,739
    What's not covered is most of the preventative things.

    79
    00:07:39,740 --> 00:07:42,290
    And I'll give you a perfect personal example.

    80
    00:07:42,770 --> 00:07:48,950
    My daughter got married not long ago to a fine gentleman whose father died in the Detroit airport when he was six.

    81
    00:07:49,430 --> 00:07:58,470
    He had sudden cardiac death. So he's now my son in law, and I know enough about cardiac disease.

    82
    00:07:58,490 --> 00:08:02,750
    I was worried that he might have an elevated value of certain lipoprotein,

    83
    00:08:02,750 --> 00:08:08,810
    called lipoprotein little A, which is responsible for sudden cardiac death in young people.

    84
    00:08:09,680 --> 00:08:13,130
    So, the value that's normal goes up to 30.

    85
    00:08:13,700 --> 00:08:23,840
    And so when Nick approached his doctor and said, can I get this test not covered because he didn't have any cardiac symptoms, so I paid for the test.

    86
    00:08:24,020 --> 00:08:32,540
    His lipoprotein level was 122, and I want my grandchildren to have a dad, and not till age six only.

    87
    00:08:32,990 --> 00:08:39,680
    So what I've learned through this is that we ultimately have to pay far more attention to the preventive part,

    88
    00:08:39,800 --> 00:08:43,850
    because it'll save us a fortune on the other end and greatly improve care.

    89
    00:08:44,090 --> 00:08:51,229
    So that's where it is right now. I think insurance companies are going to realize that paying $30 for a lipoprotein

    90
    00:08:51,230 --> 00:08:57,350
    little A test is way to their benefit... than letting the sudden cardiac death occur.

    91
    00:08:57,800 --> 00:09:01,640
    That's a very powerful example. And so thanks for sharing that experience.

    92
    00:09:02,240 --> 00:09:06,770
    Well, uh, Doctor Laposata, it was very inspiring to have you on the pulse today.

    93
    00:09:07,100 --> 00:09:11,960
    And I'll. Sure, we'll follow up with you more about exciting news from the pathology department at UTMB.

    94
    00:09:12,260 --> 00:09:15,980
    Thanks, Doctor Reiser. Thank you for tuning in to the pulse.

    95
    00:09:16,460 --> 00:09:21,710
    We hope you've gained valuable insights, discover new perspectives, and feel inspired by the stories shared today.

    96
    00:09:23,000 --> 00:09:27,830
    Please join us for future episodes to stay informed and connected right here on The Pulse.

More Information

 

Dr. Michael Laposata is a leading expert in blood coagulation and laboratory medicine. He specializes in diagnosing bleeding and clotting disorders and currently focuses on evaluating children accused of child abuse. His work has led to the reunification of over 40 families by identifying misdiagnosed medical conditions. Dr. Laposata has published over 190 peer-reviewed articles and edited nine books. He is recognized for his research in fatty acids and diagnostic errors, contributing to the National Academy of Medicine’s report on improving diagnosis in healthcare. He also oversees UTMB’s clinical laboratories, specializing in complex diagnostic cases.

For a full bibliography of Dr. Laposata’s publications, visit the NIH website.