"18ft-wide passenger friendly cabins" sounds like a line straight out of some 1970s era Delta or United ad. The truth is the 767 cross-section is pretty damn friendly, as only one poor bastard per row has to suffer a middle seat.
Continental reserves a seat in the BusinessFirst cabin (7D) on all long-haul flights for crew rest. We take turns catching some shut-eye in it on a predetermined schedule on flights with additional crew members. The 767 economical cruise speed is .1-.2 Mach less than a DC-10 operating under similar conditions, but the fuel burn savings is significant, I would guess anywhere between 12-30% per sector based on a number of factors. As far as passenger capacity, the two aircraft are very similar. In Continental's Atlantic DC-10 configuration, we fit 38 BusinessFirst seats in with 204 Coach. The 764 in Atlantic configuration fits 35 BusinessFirst and 200 Coach, a loss of 3 premium and 4 coach seats. Our DC-10 Pacific configuration had 28 BusinessFirst and 254 Coach seats, while our 767 Pacific set has 20 BusinessFirst and 235 Coach, for a total of 8 premium and 19 coach seats lost. Capacity-wise, the 767-400 is very close to the DC-10.
A 777-100 wouldn't work because operating costs would remain very similar to its big brother, this combined with a large loss of seats to counteract the higher operating cost would make the cost of incorporating a new fleet outweigh any possible benefit. Stretches work, we've seen it over and over again, but in the airline industry, very rarely does a shortened plane pan out unless substantial weight savings can be gained (of course exceptions such as A319, 737-500, and A330-200 exist) . That's why the 747SP, A310, L1011-500, and 737-600 were either outright failures or never attracted the orders the manufacturers anticipated. Stretches work better because the ratio of revenue opportunities to operating cost is improved. The 767-300, A340-300, 737-800, 727-200, and DC-8-50/60 all were commercial success stories.