Medical Discoveries in the 21st Century: An Overview

The first quarter of the 21st century has produced a striking concentration of medical breakthroughs, from gene-editing tools that can correct inherited diseases at their source to artificial intelligence systems that predict protein structures in seconds. Some of these advances have already changed clinical practice; others remain in early trials but have demonstrated results that would have seemed implausible a generation ago. What follows is a survey of the most consequential areas of progress, chosen not for their novelty alone but for their demonstrated ability to treat, diagnose, or fundamentally reframe how medicine works.

Gene Editing Moves From Lab to Clinic

CRISPR-Cas9 gene editing has arguably been the most transformative single technology to emerge in 21st-century medicine. The tool allows researchers to cut DNA at a precise location and either disable a harmful gene or insert a corrected one. Its clinical impact arrived fastest in blood disorders. In sickle cell disease, a condition caused by a single mutation in the hemoglobin gene, CRISPR-based therapies edit a patient’s own blood-forming stem cells outside the body and then transplant them back. This approach potentially offers a permanent cure regardless of whether a suitable bone marrow donor exists.

Early clinical results have been remarkable. In a trial reported in the New England Journal of Medicine, both a sickle cell patient and a patient with beta-thalassemia showed high levels of gene editing in their bone marrow and blood more than a year after treatment, along with increases in fetal hemoglobin and freedom from transfusions. The sickle cell patient experienced no further pain crises.¹ A separate study using a different CRISPR approach followed three sickle cell patients for six to eighteen months and found stable fetal hemoglobin induction, with fetal hemoglobin reaching roughly 19 to 27 percent of total hemoglobin and symptoms of the disease declining during follow-up.² In 2023, regulatory agencies in the UK and the US approved the first CRISPR-based therapy for sickle cell disease, making it the first approved treatment built on gene editing.

Beyond blood disorders, CRISPR has opened doors in areas that were practically untouchable before. Researchers are exploring its potential in muscular dystrophy, certain inherited forms of blindness, and even some cancers. Meanwhile, the broader genomics revolution has been accelerated by next-generation sequencing technologies, which have made it possible to read a patient’s entire genome quickly and affordably enough to guide treatment decisions. This has fueled the rise of personalized medicine, where drug choices and dosing can be tailored to an individual’s genetic profile.³

mRNA Vaccines and the Platform Behind Them

The COVID-19 pandemic compressed what would normally be a decade of vaccine development into about a year. But the speed was possible only because of foundational work on messenger RNA (mRNA) technology that had been ongoing since the early 2000s. The idea is elegantly simple: deliver a snippet of genetic instructions into a person’s cells so those cells temporarily produce a viral protein, which the immune system then learns to recognize and attack. The key engineering challenge was protecting the fragile mRNA molecule long enough for it to reach the right cells. That problem was solved with lipid nanoparticles, tiny fat-based capsules that shield the mRNA from degradation and help it enter cells.⁴

The success of the COVID-19 vaccines validated the mRNA platform in dramatic fashion, but the real prize may be what comes next. The same basic approach is now being tested in vaccines for influenza, respiratory syncytial virus (RSV), HIV, and several cancers. Personalized cancer vaccines, which use mRNA encoding proteins specific to an individual patient’s tumor, are already in late-stage clinical trials. The platform’s flexibility is its greatest strength: once you have a reliable delivery system, swapping in new mRNA sequences to target a different disease is comparatively straightforward.

Cancer Immunotherapy

Teaching the immune system to attack cancer has been one of the longest-running goals in oncology, and the 21st century has finally delivered therapies that work. Two broad strategies have reached clinical use. The first is immune checkpoint inhibitors, drugs that release the brakes cancer cells place on immune activity. The discovery of checkpoint proteins like PD-1, PD-L1, and CTLA-4 led to a new class of antibody drugs that have been used successfully in metastatic melanoma, kidney cancer, head and neck cancers, and non-small cell lung cancer. The FDA has approved several checkpoint inhibitors across these three categories.⁵

The second strategy, CAR-T cell therapy, is even more personalized. A patient’s own immune cells are removed, genetically reprogrammed in the lab to recognize a specific marker on cancer cells, and then infused back. In B-cell acute lymphoblastic leukemia, CD19-targeted CAR-T therapy has produced high rates of complete remission in patients who had no other curative options, though more than half eventually relapse.⁶ Newer versions are expanding the approach into T-cell cancers, which are harder to target because the therapy itself uses T cells. A recent study of CD7-targeted CAR-T therapy in patients with relapsed or refractory T-cell leukemia and lymphoma found that about two-thirds of patients responded, with manageable side effects.⁷

Relapse and side effects remain the central challenges. Cytokine release syndrome, an overreaction of the immune system, occurs in most CAR-T patients, though it is usually manageable. And the fact that many patients eventually relapse has pushed researchers toward combination approaches, using CAR-T cells alongside checkpoint inhibitors or engineering the cells to be more persistent.

GLP-1 Drugs and Their Surprising Reach

Semaglutide and liraglutide were originally developed for type 2 diabetes, but their ability to produce significant and sustained weight loss led to FDA approval as obesity treatments.⁸ What caught the medical community off guard was the cardiovascular benefit. In the landmark SELECT trial, semaglutide reduced the rate of major cardiovascular events by about 20 percent compared with placebo in people with obesity but without diabetes.⁹

A meta-analysis pooling four randomized trials confirmed the pattern, finding that semaglutide reduced major adverse cardiovascular events by about 19 percent overall. When individual outcomes were examined, cardiovascular death and non-fatal heart attacks both declined significantly, though the effect on stroke was less clear.¹⁰ These findings have turned GLP-1 receptor agonists from diabetes drugs into potentially the most important new class of cardiovascular medications in decades. Ongoing trials are examining whether the benefits extend to heart failure, kidney disease, and even neurodegenerative conditions.

Artificial Intelligence in Diagnosis and Drug Discovery

Deep learning algorithms have reached a level of diagnostic accuracy in medical imaging that, in some cases, matches or exceeds experienced clinicians. A systematic review found that in ophthalmology, AI models achieved near-perfect accuracy for detecting diabetic retinopathy, macular degeneration, and glaucoma from retinal scans. In respiratory imaging, accuracy for identifying lung nodules or lung cancer on chest scans was also high, and in breast imaging, AI showed strong performance across mammograms, ultrasound, and MRI.¹¹ These systems are not replacing radiologists or ophthalmologists, but they are being deployed as a second set of eyes, flagging cases that might be missed in high-volume clinical settings.

On the drug discovery side, the impact has been equally dramatic but in a different way. AlphaFold, developed by Google DeepMind, tackled one of biology’s hardest problems: predicting the three-dimensional shape of a protein from its genetic sequence. Determining protein structures used to take months or years using laboratory techniques. AlphaFold 2 achieved previously unseen accuracy in this task in 2020, and in 2021 the AlphaFold Protein Structure Database was released, making hundreds of millions of predicted structures available to researchers worldwide.¹² AlphaFold 3, released in 2024, expanded beyond proteins to predict how proteins interact with DNA, RNA, and small molecules, a capability that accelerates drug design by showing researchers how potential drugs might fit into their targets.¹³ Accurate protein structure prediction is already reshaping the early stages of the drug discovery pipeline, making previously inaccessible targets available for therapeutic design.¹⁴

Brain-Computer Interfaces and Neuromodulation

Brain-computer interfaces (BCIs) have made the leap from laboratory demonstration to daily, independent use. A recent study reported that a man paralyzed by ALS used an implanted brain-computer interface at home, with no researchers present, for nearly two years. He communicated over 1.9 million words across more than 183,000 sentences at an average rate of 56 words per minute. When formally tested, attempted speech was decoded with greater than 99 percent word accuracy from a vocabulary of 125,000 words. He also used the system as keyboard and mouse input for his personal computer, enabling him to send emails, browse the internet, and maintain full-time employment.¹⁵ That level of real-world, sustained use represents a critical step toward BCIs becoming practical assistive technology rather than research curiosities.

Deep brain stimulation (DBS), which involves surgically implanting electrodes in specific brain regions, has become an established therapy for Parkinson’s disease and is being explored for treatment-resistant depression. In Parkinson’s patients, DBS of the subthalamic nucleus has produced significant improvement in motor symptoms and early improvement in depression scores, though the mood benefits tend to fade over the long term.¹⁶ ¹⁷ For treatment-resistant depression specifically, preliminary DBS trials have shown promising results, though the field is still working out which brain targets produce the most durable responses.¹⁸

Regenerative Medicine and Xenotransplantation

Induced pluripotent stem cells (iPSCs), created by reprogramming adult cells back to a state where they can become almost any cell type, have gone from a theoretical concept to clinical trials in about fifteen years. iPSC technology has found applications in disease modeling, drug screening, and, most ambitiously, cell replacement therapy.¹⁹ The first clinical use of iPSC-derived cells occurred in Japan in 2013, when Masayo Takahashi’s team at the RIKEN institute transplanted iPSC-derived retinal cells into a patient with age-related macular degeneration.²⁰ Since then, clinical trials have expanded into Parkinson’s disease, where iPSC-derived dopamine-producing neurons are being tested as replacements for the cells that degenerate during the disease.

Xenotransplantation, the transplantation of animal organs into humans, has also taken a major step forward thanks to genetic engineering. The central obstacle has always been the human immune system’s violent rejection of animal tissue. Researchers have used gene editing to modify pig organs, knocking out the genes responsible for the proteins that trigger immune rejection and adding human-compatible genes.²¹ In 2022 and 2023, genetically modified pig hearts and kidneys were transplanted into human patients in closely watched procedures. The results have been mixed but instructive, with some patients surviving for months, providing data that would have been impossible to gather any other way. If the approach matures, it could dramatically ease the organ shortage crisis.

Closed-Loop Insulin Delivery

For people with type 1 diabetes, managing blood sugar has historically meant a relentless cycle of finger pricks, calculations, and insulin injections. The “artificial pancreas,” more formally known as a closed-loop insulin delivery system, automates much of that work. A continuous glucose monitor reads blood sugar levels in real time, feeds the data to an algorithm, and the algorithm directs an insulin pump to adjust delivery without the wearer needing to intervene. Randomized studies have shown that overnight closed-loop delivery increased the time blood sugar stayed in the target range by a median of 22 percent and essentially eliminated dangerously low glucose episodes after midnight.²² These benefits have held up during both nighttime and daytime use under real-world conditions.²³

Several hybrid closed-loop systems are now commercially available, and the evidence supporting their effectiveness continues to grow.²⁴ Current systems are called “hybrid” because they still require the user to announce meals and confirm bolus doses. Fully automated systems that handle meals without user input are in development but remain technically challenging because of the speed mismatch between how quickly carbohydrates hit the bloodstream and how quickly subcutaneous insulin takes effect.

Non-Invasive Prenatal Testing

Before the 2010s, screening a pregnancy for chromosomal conditions like Down syndrome relied on combinations of ultrasound and blood markers that had significant false-positive rates, often leading to invasive procedures like amniocentesis. Non-invasive prenatal testing (NIPT), which analyzes fragments of fetal DNA circulating in the mother’s blood, changed the equation. The technology is now widely established for screening for the most common chromosomal conditions, particularly trisomies 21, 18, and 13.²⁵

NIPT performs best for Down syndrome (trisomy 21), where it has high sensitivity and specificity. Its accuracy drops somewhat for trisomy 18 and more noticeably for trisomy 13 and sex chromosome conditions.²⁶ A cohort study found that the positive predictive value for trisomy 21 was about 90 percent but dropped to roughly 68 percent for trisomy 18 and 27 percent for trisomy 13, meaning a significant share of positive results for the rarer conditions turn out to be false alarms.²⁷ This is an important nuance for expecting parents: NIPT is a screening test, not a definitive diagnosis, and a positive result for a rarer condition still needs confirmation through amniocentesis or chorionic villus sampling.

Microbiome-Based Therapies

The idea of transplanting fecal matter from a healthy person into a sick one sounds crude, but fecal microbiota transplantation (FMT) has become one of the most effective treatments in gastroenterology. Its primary application is recurrent Clostridioides difficile infection, a gut infection that can become a devastating cycle of illness and antibiotic failure. Early case series reported cure rates of about 90 percent.²⁸ That figure has held up in more rigorous testing. A randomized, double-blind trial found that 90 percent of patients receiving FMT had resolution of their C. difficile infection at eight weeks, compared with 33 percent of those receiving standard vancomycin alone.²⁹

The FDA has since approved standardized microbiome products derived from donor stool, which is a significant practical step because it removes the need to find and screen individual donors. Researchers are also exploring FMT and engineered microbial communities for inflammatory bowel disease, metabolic conditions, and even certain neurological disorders, though the evidence outside of C. difficile remains much earlier-stage.³⁰

New Antibiotics Against Resistant Bacteria

Antibiotic resistance has been called one of the greatest threats to global health, and the pipeline for new antibiotics has been worryingly thin for years. One reason the problem is so acute is that gram-negative bacteria, which include many of the most dangerous hospital-acquired pathogens, have an outer membrane that acts like a fortress wall, keeping most drugs out. Darobactin, discovered in the late 2010s by screening bacteria that live symbiotically inside nematode worms, represents a genuinely new approach. It targets BamA, an essential protein in the outer membrane of gram-negative bacteria, and has shown activity against important pathogens in both laboratory experiments and animal models of infection.³¹

Darobactin is still preclinical, meaning it has not yet been tested in humans. But its discovery matters beyond the molecule itself. The finding that bacterial symbionts of animals harbor antibiotics that are particularly well suited for development into human therapeutics suggests a largely untapped source of new drugs. Most antibiotics in clinical use were originally found in soil bacteria; expanding the search to symbiotic bacteria opens a new ecological frontier at a time when the old one has been heavily mined.

RNA Interference Therapeutics

While mRNA vaccines get the headlines, a quieter revolution has been unfolding in RNA interference (RNAi) therapeutics. These drugs use small interfering RNA (siRNA) molecules to silence specific disease-causing genes after they have been transcribed but before the harmful protein is made. Several RNAi drugs have now been approved, with the highest-profile success being the treatment of transthyretin amyloidosis, a progressive disease caused by misfolded proteins that accumulate in the heart and nerves. Delivery has been the central challenge, and current approved therapies rely on lipid nanoparticles or specialized chemical tags that direct the drug to liver cells.³²

The limitation is telling: because current delivery systems work well mainly in the liver, most approved RNAi drugs target liver-expressed genes. Expanding RNAi therapy to the brain, muscles, lungs, or kidneys will require new delivery technologies. That is an active area of research, and success there would open the door to silencing disease-causing genes across a wide range of conditions that currently have no targeted treatment.

Psychedelics and Rapid-Acting Antidepressants

Depression treatment has been stuck in the same basic framework for decades: drugs that take weeks to work and fail entirely in a significant fraction of patients. Ketamine and psilocybin have disrupted that pattern by producing rapid and sometimes dramatic antidepressant effects, often within hours or days rather than weeks. They appear to work through partially overlapping but distinct brain mechanisms. Psilocybin seems to reset activity in the brain’s default mode network, a hub of self-referential thinking that tends to be overactive in depression, producing effects that may be more sustained over time. Ketamine acts more specifically on prefrontal networks involved in emotional regulation, which may explain its particular effectiveness against anhedonia, the inability to feel pleasure.³³

An interesting distinction has emerged from research into the subjective experience during treatment. A meta-analysis found that the psychological effects a person experiences during a session correlate moderately with therapeutic benefit for both drugs, but the correlation is stronger for psilocybin than for ketamine. The subjective experience accounted for roughly a quarter of psilocybin’s antidepressant effect, compared with about a tenth for ketamine.³⁴ This suggests the two drugs are not interchangeable even if they both treat depression quickly; the mechanisms and the role of the psychedelic experience itself differ. At the molecular level, researchers have identified shared immune signaling pathways that may underlie both drugs’ rapid antidepressant effects, with immune markers like IL-15 emerging as potential hubs.³⁵ Esketamine, a nasal spray form of ketamine, has already been approved for treatment-resistant depression in several countries, and psilocybin is in large-scale phase III trials.

Senolytics and the Biology of Aging

One of the more speculative but potentially far-reaching frontiers is the direct targeting of biological aging. As the body ages, some cells enter a state called senescence, where they stop dividing but remain metabolically active, secreting inflammatory molecules that damage surrounding tissue. Senolytic drugs are designed to selectively clear these cells. Preclinical work in animal models found that removing senescent cells delayed multiple age-related diseases, including bone loss, cardiovascular dysfunction, and neurodegeneration. That body of evidence was strong enough to prompt human trials. A phase 2 randomized controlled trial tested intermittent doses of dasatinib plus quercetin, a senolytic combination, in postmenopausal women to evaluate the effects on bone metabolism.³⁶ The field is young. No senolytic has been approved for any age-related condition, and whether clearing senescent cells in humans produces the same sweeping benefits seen in mice remains genuinely uncertain. But the fact that we now have a concrete molecular strategy for intervening in aging itself, rather than treating its individual downstream diseases one by one, represents a conceptual shift that would have seemed fringe even ten years ago.

1
PubMed. CRISPR-Cas9 Gene Editing for Sickle Cell Disease and β-Thalassemia
2
PubMed Central. CRISPR-Cas9 Editing of the HBG1 and HBG2 Promoters to Treat Sickle Cell Disease
3
PubMed Central. The use of next-generation sequencing in personalized medicine
4
PubMed Central. Lipid Nanoparticle-Based Delivery System—A Competing Place for mRNA Vaccines
5
PubMed Central. Immune Checkpoint Inhibitors in Cancer Therapy
6
PubMed Central. A Review of Clinical Outcomes of CAR T-Cell Therapies for B-Acute Lymphoblastic Leukemia
7
PubMed. Clinical Outcomes of CD7 CAR-T Cell Therapy in Relapsed or Refractory T-Cell Acute Lymphoblastic Leukemia and Lymphoblastic Lymphoma Patients
8
PubMed Central. Effect of GLP-1 receptor agonists on weight and cardiovascular outcomes: A review
9
PubMed. Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes
10
PubMed Central. Cardiovascular benefits of semaglutide: a systematic review and meta-analysis of randomized controlled trials
11
PubMed. Diagnostic accuracy of deep learning in medical imaging: a systematic review and meta-analysis
12
PubMed. The impact of AlphaFold Protein Structure Database on the fields of life sciences
13
PubMed Central. Review of AlphaFold 3: Transformative Advances in Drug Design and Therapeutics
14
PubMed Central. AlphaFold2 protein structure prediction: Implications for drug discovery
15
PubMed Central. Long-term independent use of an intracortical brain-computer interface for speech and cursor control
16
PubMed Central. Mood Effects After Deep Brain Stimulation for Parkinson’s Disease: An Update
17
Parkinsonism & Related Disorders. Long-term effects of bilateral deep brain stimulation of the subthalamic nucleus on depression in patients with Parkinson’s disease
18
PubMed Central. Deep brain stimulation in the treatment of depression
19
PubMed. iPSC-based cell replacement therapy: from basic research to clinical application
20
Cell Stem Cell. Pluripotent stem-cell-derived therapies in clinical trial: A 2025 update
21
PubMed Central. Genetically engineered pigs for xenotransplantation: Hopes and challenges
22
PubMed Central. Overnight closed loop insulin delivery (artificial pancreas) in adults with type 1 diabetes: crossover randomised controlled studies
23
Heliyon. An overview of advancements in closed-loop artificial pancreas system
24
PubMed Central. Closed-Loop Insulin Delivery Systems: Past, Present, and Future Directions
25
PubMed Central. Cell-Free Fetal DNA and Non-Invasive Prenatal Diagnosis of Chromosomopathies and Pediatric Monogenic Diseases
26
The Obstetrician & Gynaecologist. Cell‐free fetal DNA‐based noninvasive prenatal testing of aneuploidy
27
PubMed. Evaluation of the Z-score accuracy of noninvasive prenatal testing for trisomies 21, 18 and 13
28
PubMed Central. Treating Clostridium difficile infection with fecal microbiota transplantation
29
The Lancet Gastroenterology & Hepatology. Faecal microbiota transplantation versus placebo for first or second Clostridioides difficile infection (EarlyFMT)
30
Frontiers in Microbiomes. Recent advances in fecal microbiota transplantation for Clostridium difficile infection-associated diarrhea after kidney transplantation
31
PubMed Central. A new antibiotic selectively kills Gram-negative pathogens
32
PubMed Central. siRNA Therapeutics for the Treatment of Hereditary Diseases and Other Conditions
33
PubMed Central. Psychedelics and ketamine/esketamine in depressive disorders: biological mechanisms and associated neuroimaging and clinical changes
34
npj Mental Health Research. Meta-correlation of the effect of ketamine and psilocybin induced subjective effects on therapeutic outcome
35
Molecular Psychiatry. Convergent neuroimmune signaling underlying rapid antidepressant response to ketamine and psychedelics
36
Nature Medicine. Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women