Сергей Пузанов – Multiple Sclerosis. Everything you need to know (страница 10)
It wasn't until the late 1960s that the first genuine biomarker emerged: "oligoclonal bands" in the cerebrospinal fluid. Spotting OCBs offered the first concrete proof of a specific, localized inflammation within the CNS linked to multiple sclerosis. Before this breakthrough, the diagnosis was entirely "clinical"—meaning it was born strictly out of the doctor's office. Once science gained the ability to isolate intrathecal (intra-brain) antibody synthesis from the body's general immune response, MS diagnostics took a massive leap forward—oligoclonal IgG bands were detected in 95% of patients with genuine MS.
Yet, this first attempt to lock down the system ran into two distinct brick walls. First of all, harvesting cerebrospinal fluid via a lumbar puncture is an incredibly unpleasant procedure; it isn't 100% safe and frequently brings prolonged side effects. The second problem multiplies the first: roughly 10% of spinal fluid analyses yield a false negative.
A 10% margin of error is a massive vulnerability when dealing with a life-altering diagnosis. It threw physicians right back into a corner where clinical intuition and gut feeling carried more weight than laboratory data. What do you do if you are a young doctor facing a patient whose symptoms resemble MS but don't quite fit the textbook? What if the manual says one thing, but the latest revision of clinical guidelines says another? What if the spinal tap comes back completely clean, but the clinical mapping screams sclerosis? What do you do if you are only 70% certain, but the patient demands a definitive label and refuses to leave without one? Or conversely, what if the patient views MS as a death sentence, and you can't give them a straight "yes" or "no"?
The arrival of MRI didn’t just trigger a technological revolution—it fundamentally inverted the nature of diagnostics. Suddenly, "demyelination lesions" and "multiple sclerosis" could be visualized on a screen, which drastically simplified a doctor's workflow. The journey from the first experimental brain scan in 1981 to a fully standardized framework took exactly twenty years: in 2001, the McDonald criteria were adopted, legally equating the appearance of white spots on a tomograph to actual, physical clinical attacks.
On December 29, 1972, the captain and first officer of Eastern Airlines Flight EA-401 spent nine consecutive minutes trying to figure out why a tiny green nose-gear indicator light wouldn’t turn on. They yelled at each other, twisted the lightbulb, argued, called maintenance, and at some point, someone accidentally bumped the yoke. This bumped the autopilot out of altitude-hold mode, quietly drifting the aircraft off course. The plane crashed straight into the Florida Everglades, killing 101 out of 176 people on board—solely because the crew fixated on a single blinking bulb and trusted their automated instruments over reality. The cockpit of an aircraft that spent 9 minutes falling into a swamp held three highly qualified professionals, and all of them chose to trust the computer over their own senses. I urge you to look this up yourself—unfortunately, the "EA-401 crash in the Everglades" is a matter of historical fact.
Why bring this up? Automation is fantastic, but only if the system is structurally incapable of making mistakes. How often do automated systems fail? Early Tesla autoplots regularly plowed into parked vehicles. In 2003, an American Patriot missile battery operating in fully automated mode misidentified a British Tornado fighter jet as a hostile target and blew it out of the sky, killing the pilot and navigator. In 2010, high-frequency trading bots triggered a flash crash on the US stock market based on a false data loop—the market hemorrhaged billions in minutes. In 2020, Medtronic’s automated insulin delivery systems started glitching, throwing patients worldwide into severe hypo- or hyperglycemia. In every single instance, a standalone, isolated calibration error transformed into a systemic catastrophe because human operators stopped cross-checking the machines. With MS diagnostics via MRI, the situation turned out even worse: the calibration error was hardcoded into the standard operating procedure, creating a dangerous illusion of a stable and flawless diagnostic grid.
The contrast agents used in MRI scans are almost entirely built on compounds of gadolinium—a heavy, paramagnetic metal. Their actual job description has nothing to do with "detecting multiple sclerosis," "showing demyelination," or "measuring how sick a patient is." Gadolinium’s actual role is far more mundane: contrast enhancement simply maps zones where the blood-brain barrier has been structurally compromised. When the barrier inflames and turns porous—which happens during MS, brain tumors, infections, physical trauma, and an array of other pathologies—gadolinium molecules leak out into the brain tissue, making those spots burn bright on T1-weighted scans.
A structural breach of the blood-brain barrier occurs across a massive spectrum of conditions—strokes, angiomas, vasogenic edema, vasculitis, trauma, and post-traumatic inflammation. You see the exact same leaks in other autoimmune brain conditions, like systemic lupus erythematosus, sarcoidosis, and Behcet's disease. Furthermore, contrast-accumulating lesions are classic markers for radiation necrosis, post-surgical artifacts, the active phase of progressive multifocal leukoencephalopathy, and various granulomatous processes. Bottom line: bright spots on an MRI screen do not "show multiple sclerosis" and do not confirm a diagnosis. Contrast merely flags a leaky brain barrier, which can be caused by dozens of entirely distinct diseases.
The statistical probability that burning MRI lesions are genuinely tied to multiple sclerosis depends entirely on the clinical context: according to PubMed data, in specialized neurological centers auditing patients with high clinical suspicion of MS, about 50% to 70% of these hotspots are indeed the byproduct of demyelination. In the general population undergoing standard, routine scans, that figure plummets to 10–20%. This is exactly why the mere presence of lesions means nothing; their value hinges entirely on how they map against the total clinical profile.
The healthcare grid, like any massive bureaucratic machine, always moves toward equilibrium—not because it leads to objective truth, but because it streamlines standard operating procedures. When an MRI display throws up a picture full of white spots and a doctor faces a patient with vague, matching complaints, the path of least resistance is to write an equation: "burning lesions = multiple sclerosis." It’s faster, simpler, cleaner, and most importantly, it spares the physician from diving into complex differential diagnosis. Systems balance themselves through these exact shortcuts: an overworked doctor, buried under regulatory metrics and endless charts, stops hunting for absolute truth and settles for a functional compromise. The patient arrives demanding a label, and the presence of glowing spots allows the physician to stop treating MS as a rigorous "diagnosis of exclusion" and instead deploy it as a "catch-all label."
This is one of the foundational paragraphs of this book: if your diagnosis was slapped on based on a random MRI scan you were sent to after a brief, temporary bout of weird symptoms, and sclerosis hasn't bothered you once in the years since, you need to stop and think. No, your health isn't holding because "the interferon is working"—it is entirely possible that you never had multiple sclerosis in the first place. Yes, that is a quite possible—even if the diagnosis came from a doctor you trust unconditionally.
The misdiagnosis of multiple sclerosis is a shockingly common event: for example, in 2019, Marwa Kaisey and Nancy Sicotte tracked 241 labeled MS patients in the US, and discovered the diagnosis was flat-out wrong in 18% of cases—that’s 43 real people. It’s crucial to understand that the researchers only overturned a diagnosis when they were 100% certain it was completely bogus; how many borderline cases of overdiagnosis Kaisey and Sicotte missed is something we will unfortunately never know. What did the investigators expose? First and foremost, they proved that when these bogus labels were minted, the lumbar puncture was systematically ignored: it was either never performed at all, or its negative result was tossed into the trash as "non-informative."
On "Non-Informative" Lumbar Punctures
As you have already realized, multiple sclerosis is surrounded on all sides by deep-seated misconceptions, but nowhere does this diagnostic confusion manifest as clearly as in the analysis of cerebrospinal fluid. This is where the main diagnostic bottleneck lives—overgrown, tangled, and multi-layered. I will try to lay it out so that you can actually understand it, which means I will simplify a few things and use a couple of non-ideal analogies. It won't be effortless, but if you look at it closely, a lot of light will be shed on the myths surrounding the lumbar puncture.