There are two systemic ways to identify an area for which you can help develop a masterclass: 

DRG stands for Diagnosis-Related Group. It is a classification system that groups similar or closely related diagnoses and associated procedures.

Depending on how they’re counted, there are approximately 770 to 999 DRGs.

Learn more: Wikipedia – DRG

A striking fact: The top 10 DRGs account for nearly 30% of all acute hospital admissions.

The complete DRG list is publicly available:

▶ CMS DRG Index

⸻

DRG Structure: DRGs are grouped under MDCs (Major Diagnostic Categories). Each MDC typically includes both:

• Surgical DRGs: Focused on procedures
• Medical DRGs: Focused on diagnoses

Example:

MDC 08 – Diseases and Disorders of the Musculoskeletal System and Connective Tissue

⸻

Using DRGs for Expert Identification

• Surgical DRGs provide insight into commonly performed procedures. You can identify leading procedural experts for inclusion in a “Top 20” expert list.
• Medical DRGs group diagnoses. Some DRGs are broad and include dozens of specific conditions, each with its own potential expert cohort.

While expert overlap is inevitable, you’ll often find super-specialized experts - exactly the kind of specialists we want to feature.

⸻

Example: Connective Tissue Disorders (MDC 08)

▶ Direct link to Medical DRGs in Connective Tissue Disorders

This DRG includes:

• Amyloidosis
• Reiter’s disease
• Rheumatoid arthritis (adult and juvenile)
• Psoriatic arthritis

…and more.

There are renowned experts for each of these subcategories.

• Prof. Sir Mark Pepys (London) – an eminent authority on amyloidosis

▶ View interview

• Prof. Randy Cron (Alabama) – pediatric rheumatologist and expert on cytokine storm syndromes

▶ View interview

Both experts operate within the same DRG classification, yet focus on distinctly different disease entities.

⸻

The DRG system offers a structured and methodical approach to exploring medical categories and identifying expert contributors.

Use DRGs as your framework. However, it’s still important to apply your interests when choosing the specific diagnosis you’d like to explore further.

Oncology is one of the most important and widely searched by patients areas in modern medicine.

The WHO Classification of Tumors provides a well-organized, internationally accepted framework for categorizing tumors by organ system and tumor type.

These classifications serve as a solid foundation for identifying topics and experts when developing a masterclass.

While the full WHO tumor classification books aren’t free, the Table of Contents (TOC) for each volume is freely accessible. That’s all you need to get started.

▶ Browse WHO Classification of Tumors – IARC

You’ll see a grid of book covers, each corresponding to a tumor type by organ system. For example:

▶ Haematolymphoid Tumours (2024 Edition)

Click on the “Table of Contents” tab, and download the TOC PDF to see a comprehensive list of tumor types covered in that volume.

How to Use the TOC for Expert Search

The TOC provides a granular and structured list of tumors. You can use it to:

• Understand how tumors are grouped
• Identify major vs. minor diagnoses
• Generate keywords for expert searches

⸻

Example: Myeloproliferative Neoplasms

Excerpt from TOC:

• Chronic myeloid leukaemia (CML)
• Chronic neutrophilic leukaemia
• Chronic eosinophilic leukaemia
• Polycythaemia vera
• Essential thrombocythaemia
• Primary myelofibrosis
• Juvenile myelomonocytic leukaemia
• Myeloproliferative neoplasm, NOS (unclassifiable)

Within this list, three diagnoses in bold stand out as larger clinical areas:

• Chronic myeloid leukaemia (CML)
• Polycythaemia vera
• Primary myelofibrosis

Each of these diagnoses will have its own large expert community of hematologists-oncologists who truly specialize in it.

For a robust expert list, while you can look up experts in “Myeloproliferative neoplasms” and note the “top-20” that come up, but you probably also want to search individually for each major diagnosis above.

⸻

AML Example: Focus on Major Category + Mutations

In contrast, “Acute myeloid leukaemia (AML)” is subdivided in TOC by genetic mutations (see list below).

• Searching “Acute myeloid leukaemia” will give you many top experts.
• But reviewing mutation-specific subtypes can reveal researchers who discovered or specialize in certain genetic drivers of AML—valuable contacts for niche or high-level expert insights.

TOC part for AML:

Acute myeloid leukaemia with defining genetic abnormalities

• Acute promyelocytic leukaemia with PML::RARA fusion
• Acute myeloid leukaemia with RUNX1::RUNX1T1 fusion
• Acute myeloid leukaemia with CBFB::MYH11 fusion
• Acute myeloid leukaemia with DEK::NUP214 fusion
• Acute myeloid leukaemia with RBM15::MRTFA fusion
• Acute myeloid leukaemia with BCR::ABL1 fusion
• Acute myeloid leukaemia with KMT2A rearrangement
• Acute myeloid leukaemia with MECOM rearrangement
• Acute myeloid leukaemia with NUP98 rearrangement
• Acute myeloid leukaemia with NPM1 mutation
• Acute myeloid leukaemia with CEBPA mutation
• Acute myeloid leukaemia, myelodysplasia-related
• Acute myeloid leukaemia with other defined genetic alterations

Acute myeloid leukaemia defined by differentiation

• Acute myeloid leukaemia with minimal differentiation
• Acute myeloid leukaemia without maturation
• Acute myeloid leukaemia with maturation
• Acute basophilic leukaemia
• Acute myelomonocytic leukaemia
• Acute monocytic leukaemia
• Acute erythroid leukaemia
• Acute megakaryoblastic leukaemia
  

⸻

Thus, the WHO Classification of Tumors offers a systematic and scalable way to select high-impact topics for oncology masterclasses and to identify true global leaders in each tumor category.

Focus on major diagnostic entities first, then zoom in on subtypes or mutation-defined conditions when needed.

Once you’ve generated a list of keywords using DRGs or the WHO Classification of Tumors, the next step is to identify leading experts worldwide for the relevant diagnosis, procedure, or group of related conditions.

Here’s how to do it:

1. Register for a free account.

▶ https://app.dimensions.ai

• Once logged in, enter one of your keywords (for example, Polycythaemia vera) in the search bar at the top.

2. Click on the right-side tab that says “Researchers.”

Example: Polycythaemia vera – Researchers Tab
(open after logging into Dimensions AI)

2.1. Default opened tab is "Aggregated".

It gives you an excellent, publication-based list of researchers (clinical and academic) who have published the most on that topic.

Sorting by number of publications is usually best (this is the default), but you can also sort by citation count—sometimes that reshuffles the list significantly, depending on the keyword.

2.2. Use the “Network” Tab

Next to “Aggregated,” you’ll find the “Network” tab. Click it.

This generates a visual map of the top 25, 50, or 100 researchers and their co-author networks.

The size of each “bubble” reflects the number of publications, and you’ll immediately see which groups collaborate frequently and who the central figures are.

For clarity, I usually select “50 researchers”—it’s enough to highlight the major players without overcrowding the view.

Example: With Polycythaemia vera (~27,000 publications), the map will appear right away.

But: If you use a higher-volume term like Chronic myeloid leukaemia (~547,000 publications), the system will tell you to reduce the dataset to fewer than 25,000 publications to load the "Network" map.

⸻

3. How to reduce the number of publications to generate “Network” map.

To use the “Network” tab in Dimensions AI, the total number of publications in your search results must be under 25,000.

Many high-volume topics (like Chronic Myeloid Leukemia, or CML) exceed that limit. Here’s how to narrow your dataset.

There are two main ways to reduce the number of publications:

1. Limit by Year

• Use the “Publication Year” filter on the left panel.
• For example, CML generates more than 30,000 papers per year, so even selecting just 2019 may not be enough to fall under the 25,000 threshold.
• Still, this is a useful first filter and might work for moderate-volume topics.

2. Use Additional Filters

If year filtering alone doesn’t bring the publication count below 25,000, combine it with other filters:

a. Publication Type

• In the left panel, select “Publication Type” → “Article” (or “Monograph.”)
• This can trim thousands of entries, though in high-volume areas like CML, it may still leave you with too many results (e.g., ~26,000 for CML in 2023).

b. Fields of Research

Use additional filter: “Fields of Research”:

• Select “Biomedical and Clinical Sciences” to narrow scientific categories.

→ This narrows CML-related papers in 2023 to about 21,380, which will allow you to generate a Network map.

• Alternatively, choose “Oncology and Carcinogenesis”, a very specific subfield.

→ This reduces the count for CML in 2023 to just under 9,000—well below the limit.

What You’ll See

• Once your published paper numbers are below 25,000, the Network map becomes available in the Network tab.
• Even when using different filters (e.g., by research field), you’ll notice that many top researchers overlap, which makes sense given their publication activity in several categories.

The second method to identify leading academic doctors in any diagnostic group is to use PubMed:

▶ https://pubmed.ncbi.nlm.nih.gov/

Enter a keyword and you’ll get a list of papers. Let’s continue with “Polycythemia vera”.

I usually set:

• Display options to 200 per page
• Sort by “Most recent”

As of now, “Polycythemia vera” returns around 9,450 papers, and “Chronic myeloid leukemia” about 39,740 (these numbers change over time).

Next steps:

1. Narrow by Article Type
On the left side, check the “Review” filter under Article type.

• For “Polycythaemia vera”: ~1,450 reviews
• For “Chronic myeloid leukaemia”: ~5,421 reviews

2. Skim Quickly Through Titles

• Look for titles with 1–2 (or at most 3) authors
• Focus on short titles that bold your keyword
• Don’t overthink. Command-Click (or Ctrl-Click) to open promising ones in new tabs

These are more likely to be comprehensive, general reviews by experts in the field.

3. Note the Authors

• In reviews, the last author is typically the senior researcher
• Exception: in some European publications, the senior professor might be listed first
• If there are just two authors, chances are both are senior contributors

4. Look for Patterns

• You’ll begin to notice recurring names, publishing reviews on the same topic year after year.
• That’s a strong sign of an active expert in that field.

5. Explore Author Profiles

• Click an author’s name (Command-Click to open in a new tab)
• Sort their papers by “Most recent”
• Author with >200 publications is usually well-established
• <50 publications might indicate a younger, rising expert
• Many leaders in their field will have 300–500+ publications

⸻

Two Additional Tips:

(a) If the most recent publications are 5+ years old, that expert may be retired—still potentially valuable, but harder to reach.

(b) In the upper left corner, you’ll see a “Results by year” graph.

Click the expansion icon (square with arrows) to enlarge.

You’ll start to recognize patterns:

• “Old Masters” → Left-side peak, tapering in recent years
• “Active Ascenders” → Rising curve, peaking in the most recent complete year
• “Key Opinion Leaders” → Steady publication volume across 15–20 years, including recent ones.

By following this method, you can identify and assess leading researchers in any diagnostic category.

It takes a bit more effort than using other databases. But it’s far more granular and lets you explore each expert’s core themes and publication trajectory in real time.

Use several LLMs to generate a “master list” of questions tailored to a specific expert’s research and clinical focus.

1. Tools to Use

Use multiple language models to generate a comprehensive list of interview questions.

• ChatGPT (Standard or "Research")
• Perplexity (Standard or Pro)
• Gemini (Standard or Deep Research)
• Grok (Standard or Deep Research)
• DeepSeek (Standart or DeepThink)

Feel free to try other models you like as well.

2. Suggested Prompts

We will provide you with several versions of prompts (and please share any improvements you make).

3. LLM Output Tips

• Each LLM should generate a bio and a list of 50 questions, grouped by topic.
• If ChatGPT stalls after 30–40 questions, simply type “continue” to finish the list.

4. Reviewing the Outputs

• You’ll notice overlapping questions across models: that’s a good sign of relevance.
• Pick one LLM’s output as your base template.
• Many LLMs will group questions under “chapters” or topic headers. Preserve these and use your judgment to:

• Consolidate overlapping chapters
• Keep the strongest version of any repeated question.

5. Choosing What to Keep

When evaluating which questions to keep:

• Keep questions that appear across multiple model outputs.
• Add unique, high-quality questions from any LLM to your master list.
• For each question, ask:

“Is this relevant to a patient with this diagnosis?”

• If yes, keep it.
• If the question is purely academic or aimed at researchers, leave it out.

6. Renumber and Polish

Once your list is finalized (whether it’s 50 questions or not), run a clean-up prompt in your preferred LLM (prompt will be provided).

7. Final Steps

• Copy the final version of your question list draft into a Word, RTF, or other editable document.
• Email it to the expert and any resident or fellow involved in the interview prep.
• CC me on that email as well.

And that’s it! You now have a clean, organized draft of the master question list, ready for expert review.