Stephen Strum

MD, Private Practice of Hematology & Oncology
  • United States of America

Recent Comments

Sep 14, 2025

Nick, thanks for the kind words.  I believe that scientists often get pigeonholed in their thinking. We do not see the bigger picture, and the failure to develop a broader perspective, a more comprehensive view of what seems to be reality, remains underdeveloped. 

I have, for necessity's sake and the sake of what you might call "self-contentment," incorporated much philosophy into the clinical care of patients under my care.  The SAIN (Systems Analysis & Integrity Networking) is an expression of the oneness of life, what we glibly call the uni-verse or one story.  We scientists, more often than not, fail to see the ceaseless exchange between the biotic and the abiotic, between the breath of life and the quiet turning of stones, tides, and winds.  When we introduce "commonality" into the world of the MD (medical detective), we necessarily also realize the value of methodologies that see this universality.  This is why arenas such as gene expression profiling (GEP), 'omics, and technologies like OLINK should be employed in the biological assessment of our status. 

Seventeen years ago, I was asked to present this approach to a group of physicians at a conference on Anti-Aging Therapeutics.  I agreed to do so despite knowing that some presentations I considered unacceptable based on inadequate validation.  The positive impact of my decision to present this topic was that it prompted me to put this approach into writing.  I do not know if this Research Community accepts links to documents but this url allows you and others to get a more comprehensive picture of what I have used in real-world medical practice- and it works to the benefit of the patient. 

https://www.dropbox.com/scl/fi/07tulwyw98zgnaahde0u9/Strum-2008-SAIN-A4M-Presentation.pdf?rlkey=ny57jaoc29vdsabgwfsfw0ttj&dl=0

 

Sep 14, 2025

Continuation of the comment below. 

That study published in 1971 involved 280 patients diagnosed at the  U of Chicago btween 1931 and 1964, with 40 patients surviving ≥ 10 years.   

Of the forty patients, twenty-nine have died. Eighteen of these patients· had postmortem examinations. In sixteen, histologic evidence of Hodgkin's disease was observed.  

In the group of twenty-nine patients, the cause of death could be determined in twenty-one. Hodgkin's disease could be implicated as a significant contributory cause of death in fourteen. In seven patients, the cause of death could not be directly attributed to Hodgkin's disease but to causes such as disseminated breast carcinoma, malignant melanoma, coronary occlusion, enteritis, nephritis, and hepatitis secondary to radiation, pneumonia secondary to chemotherapy, and pulmonary fibrosis secondary to. radiation, and systemic coccidioidomycosis.  In six of these, autopsies were performed, and in four, Hodgkin's disease was present.  In the patients with systemic coccidioidomycosis and pulmonary fibrosis, no histologic evidence of Hodgkin's disease was found.

The conclusion presented in that publication from 54 years ago was that "The finding of persistent Hodgkin's disease in long-term survivors, especially in those dying from apparently unrelated causes, suggests that in some patients with Hodgkin's disease, clinical cure may in fact represent a state of equilibrium in which the host has come to terms with his disease."

My focus in oncology changed to prostate cancer in the early 1980s as a result of my collaboration with Fernand Labrie in treating metastatic disease with combination anti-androgen receptor blockade + LHRH-agonist (LHRH-A) therapy. Many of these patients went into a profound remission with a significant reduction in tumor burden based on a sustained decrease in ultrasensitive PSA levels to undetectable levels.  Moreover, an appreciation of the many fuels used to sustain cancer growth allowed for interventions that were and continue to be an integrative approach to cancer patient management: 

Cancer Fuels
1. High glycemic carbohydrates [check 2hr-PPBG] i.e., PPBG (post-prandial blood glucose) 
2. LDL [check NMR LipoProfile] via LipoScience at http://www.lipoprofile.com/index.cfm 
3. Omega 6 fatty acids [check Comprehensive Fatty Acid [CFA] profile] via Mayo Medical Labs at http://bit.ly/1AtZ7 
4. Arginine, Copper [review supplements, diet]
5. Glutamates, aspartates, cysteine [review diet]
6. pro-inflammatory cytokines [check IL-1β, TNF-α, PLA2, Rx COX2 inhibitors]
7. bone-derived growth factors (BDGF) [check bone resorption markers  DpD (deoxypyridinoline) on urine, CTX (C-Terminal Telopeptide, b-Crosslaps) on serum, TRAP5β only available in Europe
8. Hypoxia
9. Hypercoagulability [check coagulation panel], e.g.,  PT, PTT, fibrinogen, fibrin monomer, anti-thrombin III, platelet count

All of these, and likely more unidentified energy sources for the cancer cell population, became part of what was termed the SAIN (Systems Analysis & Integrity Networking) methodology in treating those with neoplastic disease.  This multi-factorial view of the tumor microenvironment (TME) has allowed the prolonged survival of many men with prostate cancer.  I would expect that if detailed autopsies were performed on this population, microscopic or very low tumor burdens would be found, similar to those in the HD study from 54 years ago. 

 

Sep 14, 2025

I am a cancer doc, a medical oncologist of 63 years clinical experience with my early years spent working on immune tolerance in neonatal mice and then onto hematopathology of HD (Hodgkin’s disease)- all at the University of Chicago. In 1971, I co-authored a paper: 

Strum SB, Rappaport H: The persistence of Hodgkin's disease in long-term survivors. Am J Med 1971, 51(2):222-240.

In all of the patients with HD that appeared cured or at the very least were in clinical remission long-term, we found microscopic evidence of persistent HD. 

I will continue this post later today. Just got called to an emergency- sorry. 

Stephen B. Strum, MD, FACP 

Sep 14, 2025

I am what you might call a "duffer" in comparison to Dr. Ahmed.  I am a hematologist/oncologist who will turn 83 next week. I have spent 63 years in the medical field, encompassing both research (animal and human) and real-world clinical care of thousands of cancer patients, from 1970 to the present.  A significant focus of my work involves prostate cancer. 

I was open to various forms of data analysis after reading about nomograms that correlated clinical and pathological findings with surgical outcomes.  I was surprised, in a bad way, at the resistance I encountered from my medical "colleagues."  This close-mindedness (my diagnosis) was mimicked ten years later with the introduction of artificial neural nets (ANNs).  And now, I am seeing a similar reaction to the use of large language models (LLMs) that employ AI assistants.  

I am on my 8th AI assistant. Initially impressed by using this for enhancing knowledge about particular medical issues,  my admiration waned when I encountered the errors, euphemistically termed "hallucinations."  I have been shocked that LLMs in the marketplace are not programmed to access public sites such as PubMed, Google Scholar, and perhaps others like ResearchGate or Sci-Med.  When I saw the fabricated medical citations in almost every AI assistant, I was in disbelief.  I found AI assistants that fabricated/hallucinated some classical works in poetry. But I was impressed that for the latter, the AI did provide some new and provocative thoughts.  

What I see is a mixed bag of a new technology that has the same caveats as existing computer software. These are, in my opinion, problematic in that they relate to the User-AI interaction that is, in theory, supposed to improve efficiency and quality of gathered information. Much of the problem, I believe, refers to programmers whose reality exists within a bubble that is often isolated from the bubble of clinicians like myself, who look to solve problems and make lives easier —our patients' lives and our own.  The "myths" I have so far encountered with AI relate to some harsh realities or commandments: 

▶︎ Thou shalt validate data retrieved by the AI with some form of interval spot checking. 

▶︎ Thou shalt request of programmers a much more efficient interface to provide feedback with more than a thumbs up or down. I suggest, for starters, a Google-like 5-star rating along with a narrative box for text. 

▶︎ LLMs should have access to public internet sites and in more advanced models, to private sites that the User subscribes to.  LLMs are in all practicality a huge data-crunching machine. If so, then sites like PubMed, Google Scholar, and others I am ignorant of should be part of the database. This would diminish the AI from its Dr. Jekyl nature to hallucinate. 

Stephen B. Strum, MD, FACP 

Sep 11, 2025

I have just submitted my CV and a list of topics related to imaging, as seen through the eyes of a medical oncologist with a long-standing interest in that field. One issue that concerns me about being a reviewer is the time devoted to a review.  I find myself frequently having to download many of the references in the submitted manuscript. Sometimes, the reason is a lack of certainty that the authors have accurately quoted the reference in question. At other times, it relates to my own desire to know more.  I hope that with my other obligations, I can still perform at least one solid review per month.  What I believe would be helpful for reviewers like myself is for journals to offer CME credits for such work. A challenging issue in this regard is that I often encounter publications where I wonder, "did the reviewer actually spend time doing his job?"