Your left knee: from MRI to marathon

The good news comes first. Nothing inside the knee is torn or loose: the menisci, ligaments, the cartilage of the main joint, bone and fluid are all normal. Every finding is about the kneecap and its two tendons: how the kneecap runs in its groove, and how hard it presses on the way.

Lateral subluxation, 0.8 cm

With the leg straight and relaxed in the scanner, your kneecap sat about 8 mm toward the outside of the groove it should run in.

TT-TG 1.7 cm

The bump on your shin where the patellar tendon attaches sits 17 mm to the outside of the groove's centre. That's the upper limit of normal (above ~20 mm is called abnormal). Every time your thigh muscle pulls, this offset tugs the kneecap outward.

Trace chondromalacia, lateral facet

The cartilage on the back of the kneecap is mildly softened in several spots, with one tiny full-thickness crack and some bone irritation under it, all on the outer facet. That's where an outward-riding kneecap presses hardest.

Quadriceps and patellar tendinosis

Both tendons are mildly irritated and thickened from load, without any tear. This is common in runners and responds to progressive loading.

Medial meniscus extrusion, small Baker's cyst

The inner meniscus bulges slightly past the joint edge without a tear. The small fluid pocket behind the knee is usually harmless and often follows joint irritation.

Put together, this pattern is usually called patellofemoral pain with maltracking. The bench below is a working model of it: the kneecap is a sled in a groove, pulled by the quadriceps and tied to the shin. It runs walking, running, stairs and cycling, travels to 120° of bend, and shows what rehab can and can't change. The rest of the page runs the same model through a running stride, your bike, the home exercises and the gym, then covers the cartilage crack and the December marathon.

What the bench shows when you walk and run

• Walking loads the kneecap lightly (about 0.6× body weight). But the knee is nearly straight for most of the step, so the groove barely holds the kneecap and it rides outward by about 8 mm. That shallow, off-centre ride can produce the clicking, catching and "giving" feelings people often describe as locking.
• Running is the opposite. At about 40° of bend, when your foot is under you, the groove has the kneecap seated, but the joint carries about 4.4× body weight. Most of that lands on the outer facet, where the cartilage crack is: the model peaks near 11 MPa. The same moment loads the patellar tendon at about 5× body weight, which is why both tendons are irritated.
• Stairs sit in between, at about 55° and 2.3× body weight.

Rehab can't move bone, so TT-TG stays 17 mm. It changes three things on the bench. Stronger, better-timed quadriceps stiffen the inner spring. Hip strength stops the knee caving inward under load, which otherwise adds an extra few millimetres of outward pull at the anchor. A slightly quicker running cadence lowers the peak load on each step. Toggle After rehab to see all three together.

Inside one running stride

Here's a running stride on a treadmill, with your left knee's anatomy underneath following each moment. Use the slider to stop on any stage. The ones to look at are loading (just after the foot lands), where the tendons and the outer facet turn red, and swing, where the knee folds deep but carries almost nothing. Switch to Walking to see the opposite: little load, but the kneecap riding furthest outward in the shallow groove.

Why cycling hurt after a while

Per pedal stroke, cycling is easy on the kneecap. With a well-set saddle the hardest push happens at about 90° of bend, where the kneecap sits deep in its groove and presses on its largest contact patch. The model puts the outer-facet peak near 3.3 MPa, against about 11 MPa per running stride. The kneecap also hardly shifts outward, because it never gets near the shallow end of the groove.

So why did it start to hurt after some distance? Three things add up:

• The count. At 85 rpm each knee does about 5,100 strokes an hour. Each one is gentle, but an irritable facet and two irritated tendons respond to the total, and that pain typically arrives late in a ride or after it.
• The set-up. A saddle a few cm too low, or too far forward, plus a big gear or a climb at low cadence, means deeper bends and more torque per stroke. Toggle it in Fig 3: the peak rises to about 5.3 MPa, 60% more on every stroke. This is the classic cause of pain at the front of the knee in cyclists.
• Knee drift. As the hip muscles tire, the knee starts to track inward at the top of the stroke. Cleats rotated wrongly, or too little float, force the same thing. On the bench that is the anchor slide moving outward: more outward pull on the kneecap.

Six home exercises, and what each one changes

They're ordered from least to most kneecap load. The readout under each figure gives the model's outer-facet stress, and all of them sit well below the ~11 MPa a running stride puts on the same spot. That your pain has dropped since you started working out is the best sign there is: it means the knee is responding to load. Keep these going right up to race week. Doses are under each figure.

Gym machines and lifts

The same model, run on the machines you asked about. One rule explains almost all of it. A machine that pushes on your shin (open chain) loads the kneecap most near a straight knee. A movement where your foot is on the floor or a plate (closed chain) loads it most when you're deeply bent. So limit leg extensions at the top of their range, and limit squats, presses and lunges at the bottom.

Everything side by side

The model's peak stress on the outer facet of the kneecap (where the crack is) for each activity, with the knee angle where it peaks and the patellar tendon load. The bar is scaled to a running stride. Anything well below running is a safe place to build strength. Anything near running counts toward your running load for the week.

The cartilage crack: is it dangerous, and how does it heal?

The report says "a tiny linear focal full-thickness cartilage denuding and underlying subchondral bony edema overlying the lateral patellar facet". In plain words: in one small spot, a hairline line of the cartilage on the back of your kneecap has worn through to the bone, and the bone just under it is irritated, which shows up as bright "edema" on MRI. Around it the cartilage is mildly softened (chondrosis) in a few places.

Is it dangerous?

Not in an urgent sense. Small cartilage lesions like this are common on runners' MRIs, including in people with no pain, and many stay the same size for years. The real risk is slow enlargement if the same spot keeps getting overloaded. With a kneecap that rides outward, that spot is exactly where the load goes. So the goal is to calm that spot down and spread its load, not to fix the line itself.

Will it heal?

Honestly: the cartilage itself won't regrow to normal. Adult joint cartilage has no blood supply, so a full-thickness line doesn't knit back like bone or tendon. It may partly fill with tougher scar cartilage, or simply stay as it is. What does heal is the irritated bone underneath, and that's usually the main source of pain. It settles over weeks to a few months once its load is under its tolerance. Your falling pain fits that.

So what does "cured" look like?

A quiet knee. The line still shows on a scan, but it doesn't hurt, doesn't swell, and doesn't grow. You get there by lowering peak stress on the outer facet: stronger quads and hips, better tracking, a quicker running cadence, and no sudden jumps in load. Gentle, repeated loading also helps the cartilage, because it gets its nutrients from being squeezed and released. That's why cycling with a correct saddle and the exercises here are good for it, while long rest isn't.

Do I need more than exercise?

Usually not, for a lesion this small. Injections such as PRP or hyaluronic acid have mixed evidence and aren't first-line. Surgery (cartilage repair, or moving the tendon attachment to fix the tracking) is for larger lesions, repeated dislocations, or a TT-TG well above 20 mm, after months of rehab have failed. None of those describe your report. Supplements (collagen, glucosamine) are low-risk but have weak evidence.

Warning signs

The knee swells after runs (puffy, tight, hard to fully bend). It truly locks or catches and won't straighten. Pain is worse each week rather than better, or wakes you at night. Any of these means the lesion is unhappy: see your orthopaedic doctor, who may repeat the MRI.

The December marathon

You have about six weeks. Your pain is falling while you train, which is the most important fact here. Running the marathon is reasonable if three things are true: you can already run about 20 km or more without pain going above 3/10, the knee doesn't swell afterwards, and it's no worse the next morning. If one of those fails, the honest options are run-walking the whole race at an easier pace, dropping to the half, or deferring. Six weeks is too short to build both mileage and kneecap tolerance from scratch.

What the model says matters most for race day is the per-landing peak, so the cheapest wins are a quicker cadence, a run-walk rhythm from the start (walk breaks give the facet and tendon a pause before fatigue sets in and the knee starts to drift inward), and flat, even pacing, because downhills combine a deep bend with braking.

A quicker step means slightly more landings, each one lighter. Joint stress per landing matters more than the count, which is why cadence is the cheapest change you can make. Rehab strength (Fig 1, "After rehab") stacks on top of this.

When to go back to the doctor

Model sources (approximate): kneecap contact areas from Huberti & Hayes (1984); joint force from a pulley model of the quadriceps and patellar tendon with a small straight-knee offset; gait curves from standard walking and running data; cycling knee moments after Ericson (1986); open- versus closed-chain kneecap stress pattern after Steinkamp (1993) and Escamilla (1998); cadence effect from Lenhart (2014). Machine loads assume a moderate 10–12-rep-max weight.