I certainly accept that the capsules as they were designed could only make use of the minimal sort of refurbished-parts reuse that Michel has mentioned. But how hard would it be to redesign a new capsule, using a successful prior design as a benchmark, so that it can be readily disassembled in modular units, with an eye toward reuse? Apparently for instance resurfacing the exterior of the Apollo capsules with a new layer of ablative material would be a dubious thing to do, but what if the outer panels to which the ablative is bonded were themselves designed to be removed and discarded (or rather, studied then displayed in museums) so that one simply attaches new outer panels complete with ablative already built on? Similarly for landing crumple zones--design these layers to be removed and set aside as units, and bolt on another one to the inner capsule structure.
Given how many other expensive bits of hardware absolutely had to be completely discarded after one use with the basic Saturn/Apollo design, it might not be very cost-effective overall to invest in a reusable return capsule. The more hardware one can hope to recover and reuse from a given launch though, the more sense redesigning the capsule would make. It is true that many parts of it would be items that one would not want to design for an extended life (or rather, they are already overdesigned for a much longer service life than envisioned--but that is much-needed safety factor and doesn't count toward planning to actually use it; it is there to guarantee that the part lasts long enough for the projected mission). (Note though that Apollo's oxygen tanks were not part of the Command Module, but of the Service module that would certainly be discarded).
Delta Force notes that the demands on a heat shield that is merely returning from low Earth orbit are much lower (about half, overall) than those required of a shield meant for recovery from a translunar trajectory. That doesn't suggest to me though that one can safely resurface an Apollo CM that came back from Lunar orbit for one more use in LEO. Apparently NASA judged that one could not reliably resurface any capsule at all. But if this issue were addressed with my replaceable parts approach, then the same inner capsule might be purposed back and forth between translunar and LEO missions; it would then be a matter of bolting on heatshield A, a heavier item for high-energy missions, versus the lighter shield B type for on orbital mission. However I'd think the two missions would differ enough in general that one would want different controls and environmental systems anyway; it might be smarter to just design two different types of capsule.
The remark suggests the agenda the OP has might be, as in Eyes Turned Skyward, to have a standardized set of Apollo-derived vehicles, presumably launched on something in the ballpark of a Titan III or Saturn 1B (presumably in that second case improved as per ETS, replacing 8 H-1 engines with one F-1A, and the kludgy cluster tank of the Saturn 1 designs, which were lashed together from Redstone and Juno/Jupiter parts already developed before the early 60s, with a new monolithic design like a small scale version of the Saturn V first stage--the ETS Saturn 1C in fact) to use for LEO work, only going ETS one better with a reusable capsule. Designing it to take external shell replacements bolted to the inner pressure vessel might mean as much as half or more of the capsule mass (around 6 tons all up) serve many missions.
I don't know; a better approach might be to seek to make the capsule fully reusable, meaning going past ablative coatings to parts that radiate away the heat without being used up at all. That's tricky of course. The Shuttle had to go with the tiles, which we all know to our sorrow were vulnerable mechanically. Of course unlike a Shuttle, if we have a CM the main TPS is protected until the service module sloughs off. OTOH it landed by splashing down into the ocean; I don't think the Shuttle's ceramics would take kindly to that kind of treatment. We'd want something more resilient, like a metal, or else we have to arrange for another kind of landing, like the plan to make Gemini ride down on a Rogallo wing, which enables it to land sideways on skids. I believe the problem with that plan was finding a way to reliably deploy the para-wing. The three parachute solution and splashdown was redundant; one chute could fail to deploy and the other two would be adequate--but landing on a solid surface would be an emergency crash-landing that might injure or conceivably kill crew members. They never got around to perfecting the parawing. And obviously the geometry of a Gemini capsule lent itself better, a bit, than the squatter Apollo CM, to such a mode of flight.
Probably the best thing would be to press on past Apollo to a new design, say a biconic that enters sideways, with the TPS on the sloping cone surface rather than the circular "bottom"--which would be a bottom during launch of course. At entry and landing it wouldn't be. Then we'd still have to solve the parawing problem, or dunk the thing in seawater just the same, or incorporate wings in the body to make it more of a lifting body a la the winged versions of Kliper the Russians proposed last decade. Or a spaceplane design of course! Such ambitious approaches might allow reuse not only of the CM type avionics and environmental systems, but the SM type systems too--fuel tanks, oxygen tanks, the general workhorse systems that the SM contained in Apollo, and of course the orbital maneuvering engines. But that puts a lot of stuff that poses risks and threats in the same shell as the crew.
The authors of ETS probably knew what they were doing when they postponed (to past the end of their TL) reusable manned spacecraft, and accepted that each Apollo derivative craft launched would be a one-shot. In the TL was a proposal of a biconic craft such as outlined above but it was just a study.
One advantage I thought they'd reap from that would be that the equipment inside the CM (and SM) would be advancing with the general state of the art; the avionics that served in 1969 would be replaced with solid state integrated circuits that would mass a tiny fraction and draw a tiny fraction of the power. They chided me for that, arguing that the space environment is no place for microelectronics, that Apollo's mid-60s designs with their magnetic core memories and so on would be much more robust given the radiation and thermal environment. Well, I'm sure that OTL the NRO and Air Force did push for solid state equipment--of course not for manned vehicles. And their best in any given year was highly classified too, unlike the public, civil NASA craft which have their design details published freely. Still I'd think that perhaps after a conservative 1970s and '80s, microelectronics pioneered in those decades would become obsolete for spy sats and military comsat and weather sats, and be declassified (with a classified, but reassuring to NASA designers cleared to know, record of successful operation). The avionics would get better and yet lighter and less power hungry. New materials and inventions could lead to generational improvements in the other systems as well, so that a late 1990's ETS CM for an Apollo Mark V would be "the same" as the Apollo Mark II of the Lunar program in the same sense as a 1980 Volkswagen Beetle would be "the same" as a 1950 model. Probably mass a bit more, but with lots more capability and redundancy. This is one advantage of throwaway vehicles--though the constant redesigning is itself a cost. If you could justify refurbishing old capsules, you might be stuck with more legacy tech. (Which the authors of ETS figured apparently might be just fine).
I think it would be hard to put a dollar cost as you've asked for on the project of reusing Apollo CMs. With the Mark II, it would be "infinite"--no NASA team could endorse doing it at all. You'd need a new-designed Mark III. Designing it would be costly, especially as the whole point of reusability is to minimize the ongoing production costs; the contractors would be asked to make all their money producing just a half dozen or so examples, which between them might be meant to cover 60 missions or so, which would at OTL launch rates mean they last for 10-15 years. Being half rebuilt each time, but there is no point in trying unless you can get those incremental costs down to a lot less than half the cost of just building a whole new capsule. So Rockwell or whoever makes them would want to charge a very pretty penny for each example, and the cost of refurbishment parts would presumably be in the stratosphere too. They would not be set by a competitive market but by politically written contracts. And NASA would still be throwing away a new (if smaller) SM and any mission modules added, along with two or more launch stages.